Cognition and Other Predictors of Functional Disability Among Veterans With Mild Traumatic Brain Injury and Posttraumatic Stress Disorder

Abstract

Background:

Limitations in everyday functioning are frequently reported by veterans with a history of mild traumatic brain injury (mTBI) and/or posttraumatic stress disorder (PTSD). Multiple factors are associated with functional disability among veterans, including depression, poor social support, cognition, and substance use. However, the degree to which these factors, particularly cognitive capacities, contribute to functional limitations remains unclear.

Methods:

We evaluated performance on tests of processing speed, executive functioning, attention, and memory as predictors of functioning on the World Health Organization Disability Assessment Scale (WHODAS) 2.0 in 288 veterans. Participants were placed in one of the following groups: PTSD-only, mTBI-only, mTBI + PTSD, and neither PTSD nor mTBI (deployed control group). Cognitive test performances were evaluated as predictors of WHODAS 2.0 functional ratings in regression models that included demographic variables and a range of mood, behavioral health, and postconcussive symptom ratings.

Results:

Multiple cognitive test performances predicted WHODAS 2.0 scores in the deployed control group, but they generally did not predict functioning in the clinical groups when accounting for demographics, mood, behavioral health, and postconcussive symptoms.

Conclusions:

In veterans with mTBI and/or PTSD, cognitive test performances are less associated with everyday functioning than mood and postconcussive symptoms.

HUNDREDS OF THOUSANDS of US service members are estimated to have sustained a mild traumatic brain injury (mTBI) while serving in Operation Iraqi Freedom (OIF), Operation Enduring Freedom (OEF), and Operation New Dawn (OND).¹ Development of posttraumatic stress disorder (PTSD) is also common among deployed veterans, particularly those with a history of combat-related TBI, with prevalence estimates ranging from 10% to 20% or higher.^1,2 Both conditions are linked to functional limitations and disability in civilian life and numerous mental health comorbidities, including substance abuse, homelessness, and suicide.¹

Common mTBI symptoms include headache, nausea, light and sound sensitivity, vestibular dysfunction, and cognitive complaints, among others.³ These resolve in the majority of persons within 3 months,⁴ but a small proportion report persisting symptoms for 12 months or longer in both civilian and veteran populations.^1,5,6 While numerous PTSD and mTBI symptoms overlap (eg, memory and concentration problems, sleep disturbance, fatigue, headache, and vestibular),⁷ persistence of symptoms is more likely with PTSD.⁸

A significant number of service members with a history of mTBI and/or PTSD experience substantial, persisting limitations in daily living, including issues with vocational, family, social, and self-care activities.^7,9 However, the degree to which mTBI, PTSD, and other relevant variables contribute to these limitations remains unclear. The purpose of this article is to advance understanding of predictors of long-term functional disability for OIF/OEF/OND veterans with histories of mTBI alone, PTSD alone, and mTBI and PTSD, relative to deployed veterans without mTBI or PTSD. Given that cognitive changes persisting beyond 3 months are associated with limitations in everyday functioning in nonmilitary mTBI populations,^9,10 we focused specifically on the role of cognition as a predictor of functional disability.

COGNITION

The evidence regarding persisting cognitive impairment following mTBI is mixed. Multiple meta-analytic studies show that measurable cognitive impairment is common in the first 3 to 12 months following mTBI but is uncommon after 12 months.^4,5 However, other research suggests that cognitive deficits can persist in some circumstances and types of TBI.¹¹ Recent research reports long-term differences in activation patterns between mTBI and cognitively normal controls during the performance of cognitive tasks.¹² Moreover, there is increasing recognition of the heterogeneity in mTBI outcomes,¹³ with differential outcomes for deployed veterans thought to be due, in part, to injury mechanisms (ie, blast vs blunt injuries).^14,15

Performance validity testing (PVT), often referred to as “effort testing,” may help clarify the mixed findings regarding persisting (≥12 months) cognitive deficits after mTBI. Similar to reports with civilians incurring mTBI,¹⁶ research with combat veterans demonstrates that controlling for performance validity dramatically impacts neuropsychological findings. For example, one study that controlled for PVT scores¹⁷ found no differences in neuropsychological test performance in OIF/OEF veterans with and without mTBI history. Combat veterans with a history of mTBI who “failed” PVT have also been reported to score worse on neuropsychological tests than veterans with a history of severe TBI.^18,19

It has also been suggested that persisting cognitive impairments in veteran mTBI populations may be disproportionally related to psychiatric symptoms rather than brain injury itself. One OEF/OIF veteran report indicates that depression and PTSD predict neuropsychological test performance whereas mTBI does not.²⁰ Other OEF/OIF veteran studies show, similarly, that self-reported cognitive symptoms are related to depression, anxiety, and posttraumatic stress but not performance on neuropsychological tests.²¹ Furthermore, examinations of role of PTSD history, when TBI history is added to prediction models, indicate that mTBI history does not enhance predictions of neuropsychological test performance.²²

FACTORS RELATED TO FUNCTIONAL DISABILITY

In addition to cognitive deficits, other factors are thought to influence the relationship between mTBI, PTSD, and everyday functioning and disability among veterans. For example, psychiatric symptoms in a general physical injury population were stronger predictors of long-term disability than injury severity and pain symptoms.²³ In the mTBI literature, depression (in conjunction with PTSD) mediates the relationship between head injury with loss of consciousness and long-term physical health outcomes.⁷ A longitudinal study found that PTSD symptoms are stronger predictors of long-term postconcussive symptoms and functioning than mTBI history and that, after controlling for PTSD symptoms, mTBI history is not predictive of psychosocial outcomes.²⁴ In contrast, OEF/OIF/OND veterans with a history of mTBI, PTSD as well as depression, compared with veterans with PTSD and depression but no mTBI history, report greater functional disability.²⁵

CURRENT STUDY

Amongst OEF/OIF/OND veterans, mTBI, PTSD, and psychological symptoms are related to functional disability. The role, however, of persisting cognitive changes in long-term functional disability is not well understood. Accordingly, our study evaluated performance on a range of neuropsychological assessments against other potential predictors of functional disability in a population of OIF/OEF/OND veterans with histories of mTBI alone, PTSD alone, mTBI and PTSD, and deployment history but neither mTBI nor PTSD.

METHODS

Study procedures

Data were drawn from a previous diagnostic accuracy study conducted at 3 VA Polytrauma Network Sites (see Pape et al²⁶ for detailed recruitment and data collection procedures). Approval for the study was obtained from the VA Centralized Institutional Review Board and the Research and Development Committees at each site. Veterans 18 years and older who were deployed to Iraq and/or Afghanistan conflicts and had not received treatment of concussion in the preceding 30 days were eligible for enrollment. Across sites, 454 veterans were enrolled between August 1, 2010, and September 30, 2011. After excluding veterans with missing outcome data or with scores below study cutoffs on measures of symptom/performance validity, 288 veterans remained. mTBI history was confirmed with the VA’s mTBI screen and Comprehensive TBI Evaluation (CTBIE), which includes a postconcussive symptom measure, the 22-item “Neurobehavioral Symptom Inventory” (NSI).²⁶ Participants also completed a battery of measures including the World Health Organization Disability Assessment Scale (WHODAS) 2.0,²⁷ the Clinician-Administered PTSD Scale (CAPS),²⁸ various behavioral health inventories, and a range of neuropsychological tests, as outlined later.

Measures

WHODAS 2.0

WHODAS 2.0 is a general measure of functioning and disability designed to provide standardized disability ratings across populations.²⁷ The scale provides ratings for 6 life domains: cognition, mobility, self-care, getting along, life activities, and participation in society.

Behavioral health questionnaires

Behavioral health questionnaires were dichotomized on the basis of cutoff scores that suggest the presence of clinically elevated symptom levels. The Alcohol Use Disorder Identification Test–Alcohol Consumption Questions (AUDIT-C) is a 3-item self-report measure of alcohol use within the past year that assesses for hazardous drinking.²⁹ Consistent with published guidelines, cutoff scores of 4 or more (for men) or 3 or more (for women) were used. The Beck Anxiety Inventory (BAI) is a 21-item self-report measure of anxiety symptoms.^30,31 A cutoff score of 16 or more (“moderate anxiety”) was used for this measure. The Beck Depression Inventory–Second Edition (BDI-2) is a 21-item self-report measure assessing symptoms of depression.³² Here, a cutoff score of 20 or more (“moderate depression”) was employed. The CAPS²⁸ is a structured diagnostic interview based on the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) criteria for PTSD. Here, the CAPS “lenient criteria” (ie, frequency 1/intensity 2) were employed. The Insomnia Severity Index (ISI) is a 7-item self-report measure assessing severity and functional impact of sleep difficulties.³³ A cutoff score for “moderate insomnia” (>14) was used. Finally, a single dichotomous item from the CTBIE was used to assess for pain within the last 30 days.³⁴

Neuropsychological assessments

Measures of attention, processing speed, executive functioning, and verbal memory were included in the neuropsychological test battery. Baseline intellectual functioning (“IQ”) was estimated with the Wechsler Test of Adult Reading (WTAR).³⁵ The Trail Making, Verbal Fluency, and Color-Word Interference tests from the Delis-Kaplan Executive Function System (D-KEFS)³⁶ were included as measures of executive capacities such as mental flexibility, verbal fluency, and response inhibition. The omission, commission, and reaction time scores from Conners’ Continuous Performance Test– Second Edition (CPT-II)³⁷ were included as measures of attention and impulsivity. The total number of errors from the Wisconsin Card Sorting Test (WCST-64) was included as a measure of adaptive reasoning.³⁸ Total Learning across trials 1 to 5 and long-delay free recall scores from the California Verbal Learning Test–Second Edition (CVLT-II)³⁹ were included as measures of verbal learning and recall. Finally, the Processing Speed Index (PSI) from the fourth edition of the WAIS (WAIS-IV)⁴⁰ was included as a visuomotor processing speed metric.

Analyses

Demographic and clinical characteristics of the sample were described using frequencies and percentages. Neuropsychological measures were characterized using means and standard deviations consistent with the standard metrics for each test (eg, T-scores, scaled scores). Unadjusted regression models were estimated to examine the association between WHODAS 2.0 functional ratings and each demographic and clinical characteristic and each neuropsychological test score. We also estimated Pearson and Spearman correlation coefficients between the demographic and clinical characteristics and the WHODAS 2.0 functional ratings. Because the neuropsychological measures were highly correlated, and the sample size was relatively small, separate multiple regression models were estimated for each neuropsychological measure of interest. These multiple regression models included estimated IQ scores, demographics (ie, sex, race, ethnicity, education, marital status, employment, receipt of disability benefits), behavioral health ratings (insomnia, alcohol use, pain, positive TBI history), mood (anxiety, depression, PTSD), and postconcussive symptoms. Finally, the multiple regression models for each neuropsychological measure were stratified to examine the effect of that cognitive capacity on the WHODAS 2.0 functional ratings within the 3 clinical groups (mTBI-only, PTSD-only, and mTBI + PTSD) and the control group (deployed control).

RESULTS

Study sample characteristics

The demographics and clinical characteristics of the total study sample of veterans, stratified by 1 control group and 3 clinical groups (see Table 1), indicate that most participants were white (78%) males (89%) in their early 30s (33 ± 8 years) who were employed (83%) prior to deployment. While there were no differences between any of the groups for age, education, ethnicity, estimated IQ, or marital status, the PTSD-only and deployed control groups included higher percentages of women (P = .025). There were also between-group differences for both predeployment employment (P = .031) and employment at the time of study participation (P < .001). Across groups, approximately 40% of participants received some portion of their income from disability payments, with the highest rates reported from the mTBI + PTSD and PTSD-only groups (55% and 51%, respectively) and the lowest in the deployed control group (∼30%). Relative to the deployed control group, each of the 3 clinical groups had a higher mean number of self-reported postconcussive symptoms (NSI: P < .001), as well as lower WHODAS 2.0 functional ratings (P < .001). The mTBI + PTSD group had the highest mean number of self-reported NSI symptoms (17.47; SD = 4.23) and lowest level of WHODAS 2.0 function (45.28; SD = 10.38).

TABLE 1.

Demographic and clinical characteristics

	Deployed control (N = 147)	mTBI + PTSD (N = 62)	PTSD-only (N = 37)	mTBI-only (N = 42)	Total (N = 288)	P
Age, mean (SD)	33.94 (9.09)	32.01 (6.95)	33.78 (8.46)	31.16 (7.70)	33.10 (8.43)	.174 (F)
Sex, n (%)
Male	127 (86.39)	60 (96.77)	29 (78.38)	39 (92.86)	255 (88.54)	.025a (χ2)
Female	20 (13.61)	2 (3.23)	8 (21.62)	3 (7.14)	33 (11.46)
Race, n (%)
White	121 (82.31)	49 (79.03)	26 (70.27)	28 (66.67)	224 (77.78)	.111 (χ2)
Nonwhite	26 (17.69)	13 (20.97)	11 (29.73)	14 (33.33)	64 (22.22)
Ethnicity, n (%)
Hispanic or Latino	22 (14.97)	9 (14.52)	5 (13.51)	9 (21.43)	45 (15.63)	.728 (χ2)
Not Hispanic or Latino	125 (85.03)	53 (85.48)	32 (86.49)	33 (78.57)	243 (84.38)
Education, n (%)
High school or less	72 (48.98)	34 (54.84)	19 (51.35)	23 (54.76)	148 (51.39)	.843 (χ2)
Greater than high school	75 (51.02)	28 (45.16)	18 (48.65)	19 (45.24)	140 (48.61)
Marital status, n (%)
Married/partnered	62 (42.18)	29 (46.77)	12 (32.43)	21 (50.00)	124 (43.06)	.403 (χ2)
Single/divorced	85 (57.82)	33 (53.23)	25 (67.57)	21 (50.00)	164 (56.94)
Predeployment employment, n (%)
Yes	123 (83.67)	44 (70.97)	33 (89.19)	38 (90.48)	238 (82.64)	.031a (χ2)
No	24 (16.33)	18 (29.03)	4 (10.81)	4 (9.52)	50 (17.36)
Current employment, n (%)
Yes	129 (87.76)	37 (59.68)	29 (78.38)	34 (80.95)	229 (79.51)	.000b (χ2)
No	18 (12.24)	25 (40.32)	8 (21.62)	8 (19.05)	59 (20.49)
Receiving disability, n (%)
Yes	44 (29.93)	34 (54.84)	19 (51.35)	17 (40.48)	114 (39.58)	.003b (χ2)
No	103 (70.07)	28 (45.16)	18 (48.65)	25 (59.52)	174 (60.42)
Total no. symptoms (NSI), mean (SD)	7.04 (5.31)	17.47 (4.23)	13.49 (4.48)	14.17 (4.92)	11.15 (6.59)	.000b (F)
WTAR predicted WAIS-III FSIQ	104.13 (8.20)	102.61 (8.44)	100.30 (9.73)	103.19 (7.48)	103.17 (8.42)	.089 (F)
WHODAS 2.0	23.29 (13.99)	45.28 (10.38)	38.65 (10.14)	35.60 (10.65)	31.79 (15.36)	.000b (F)

Abbreviations: FSIQ, full-scale IQ; mTBI, mild traumatic brain injury; NSI, Neurobehavioral Symptom Inventory; PTSD, posttraumatic stress disorder; WAIS-III, Wechsler Adult Intelligence Scale–Third Edition; WHODAS, World Health Organization Disability Assessment Scale (lower WHODAS score indicates more function); WTAR, Wechsler Test of Adult Reading.

^aP < .05.

^bP < .01.

Mood and behavioral health scores (see Table 2) indicate that the mTBI + PTSD group reported the highest symptom rates on these measures, as well as higher rates of insomnia (P < .001) and pain (P < .001). Higher rates of depression, anxiety, and pain were reported in the clinical groups than in the deployed control group. Group means on cognitive test scores (see Table 3) suggest that the deployed control group performed better than all 3 clinical groups on most measures. There were group differences on D-KEFS Trail Making (Number Switching), Color-Word Interference (Inhibition, Inhibition/Switching), and Verbal Fluency (Category Switching Total Correct), as well as both learning and delayed recall indices on the CVLT-II and the WAIS-IV PSI. The mTBI + PTSD group generally performed the worst across all measures with significant group differences, but the mTBI-only group performed at similarly diminished levels on the D-KEFS Trail Making (Number-Letter Switching) and the WAIS-IV PSI, whereas the PTSD-only group had similarly low scores on the CVLT-II Total Trials 1–5 and Delayed Recall.

TABLE 2.

Mood and behavioral health measures

	Deployed control (N = 147)	mTBI + PTSD (N = 62)	PTSD-only (N = 37)	mTBI-only (N = 42)	Total (N = 288)	P
Mood measures						.000 (χ2)a
BAI, n (%)
None	133 (90.48)	35 (56.45)	22 (59.46)	26 (61.90)	216 (75.00)
Mild	13 (8.84)	4 (6.45)	6 (16.22)	11 (26.19)	34 (11.81)
Moderate	0 (0.00)	14 (22.58)	7 (18.92)	3 (7.14)	24 (8.33)	.
Severe	1 (0.68)	9 (14.52)	2 (5.41)	2 (4.76)	14 (4.86)
BDI-2, n (%)						.000 (χ2)a
None	130 (88.44)	14 (22.58)	17 (45.95)	22 (52.38)	183 (63.53)
Mild	14 (9.52)	16 (25.81)	10 (27.03)	11 (26.19)	51 (17.71)
Moderate	3 (2.04)	21 (33.87)	6 (16.22)	9 (21.43)	39 (13.54)
Severe	0 (0.00)	11 (17.74)	4 (10.81)	0 (0.00)	15 (5.21)
CAPS, n (%)						.000 (χ2)a
Above lenient cutoff score	0 (0.00)	62 (100.00)	37 (100.00)	0 (0.00)	99 (34.38)
Below lenient cutoff score	147 (100)	0 (0.00)	0 (0.00)	42 (100.00)	184 (63.89)
Behavioral health measures
AUDIT-C, n (%)						.412 (χ2)
Hazardous	75 (51.02)	26 (41.94)	14 (37.84)	19 (45.24)	134 (46.53)
Nonhazardous	72 (48.98)	36 (58.06)	23 (62.16)	23 (54.76)	154 (53.47)
ISI, n (%)						000 (χ2)a
Absence or subthreshold insomnia	121 (82.31)	23 (37.10)	22 (59.46)	27 (64.29)	193 (67.01)
Moderate or severe insomnia	26 (17.69)	39 (62.90)	15 (40.54)	15 (35.71)	95 (32.99)
Pain (item 17a from CTBIE), n (%)						.000 (χ2)a
No pain within 30 d	72 (48.98)	7 (11.29)	7 (18.92)	6 (14.29)	92 (31.94)
Pain within 30 d	75 (51.02)	55 (88.71)	30 (81.08)	36 (85.71)	196 (68.06)

Abbreviations: AUDIT-C, Alcohol Use Disorders Identification Test–Alcohol Consumption Questions; BAI, Beck Anxiety Inventory; BDI-2, Beck Depression Inventory–Second Edition; CAPS, Clinician-Administered PTSD Scale; CTBIE, Comprehensive TBI Evaluation; ISI, Insomnia Severity Index; mTBI, mild traumatic brain injury; PTSD, posttraumatic stress disorder; TBI, traumatic brain injury.

^aP < .01.

TABLE 3.

Neuropsychological test scores^a

	Deployed control (N = 147), mean (SD)	mTBI + PTSD (N = 62), mean (SD)	PTSD-only (N = 37), mean (SD)	mTBI-only (N = 42), mean (SD)	Total (N = 288), mean (SD)	P
D-KEFS Trail Making (ss)
Number-Letter Switching	10.37 (2.02)	9.10 (3.32)	10.68 (2.11)	9.43 (2.99)	10.00 (2.57)	.001b
D-KEFS Verbal Fluency (ss)
Letter Fluency	10.49 (3.19)	9.44 (3.10)	9.46 (3.40)	10.02 (3.51)	10.06 (3.27)	.107
Category Switching Total Correct	11.22 (3.54)	9.60 (3.75)	10.76 (3.11)	10.86 (3.20)	10.76 (3.53)	.025c
Category Switching Total Accuracy	11.59 (3.12)	10.61 (3.24)	11.22 (2.82)	11.19 (2.96)	11.27 (3.10)	.220
D-KEFS Color-Word Interference (ss)
Inhibition	11.01 (2.54)	8.97 (3.52)	10.84 (2.71)	9.81 (3.29)	10.37 (3.01)	.000b
Inhibition/Switching	10.69 (2.48)	8.85 (3.21)	9.73 (2.91)	9.67 (2.68)	10.02 (2.82)	.001b
CPT-II (T)
Omissions	46.92 (8.24)	49.95 (12.77)	46.84 (9.77)	45.85 (6.70)	47.41 (9.46)	.105
Commissions	48.10 (9.09)	51.65 (9.87)	50.51 (8.25)	49.82 (9.61)	49.23 (9.31)	.068
Hit Rate	48.73 (9.86)	47.78 (10.34)	45.73 (7.25)	47.92 (9.39)	48.02 (9.61)	.402
CVLT-II
Trials 1–5 Total (T)	54.20 (8.81)	47.31 (10.86)	50.00 (10.40)	54.29 (8.45)	52.19 (9.80)	.000b
Delayed Recall (z)	0.11 (0.95)	−0.71 (1.25)	−0.21 (1.08)	0.21 (0.91)	−0.09 (1.08)	.000b
WCST (T)
Total Errors	50.53 (8.40)	50.97 (8.33)	46.95 (8.90)	51.60 (8.13)	50.32 (8.48)	.063
WAIS-IV (SS)
Processing Speed Index	104.18 (10.27)	97.84 (13.20)	101.92 (11.56)	99.38 (10.06)	101.82 (11.36)	.001b

Abbreviations: CPT-II, Conners’ Continuous Performance Test–Second Edition; CVLT-II, California Verbal Learning Test–Second Edition; D-KEFS, Delis-Kaplan Executive Function System; mTBI, mild traumatic brain injury; PTSD, posttraumatic stress disorder; ss, scaled score (M = 10; SD = 3); SS, standard score (M = 100, SD = 15); T, T-score (M = 50, SD = 10); WAIS-IV, Wechsler Adult Intelligence Scale–Fourth Edition; WCST, Wisconsin Card Sorting Test; z, z-score (M = 0.0, SD = 1.0).

^aP values are for one-way analyses of variance.

^bP < .01.

^cP < .05.

Predictors of WHODAS 2.0 functional disability ratings across study participants

For the entire study sample, unadjusted regression results (see Supplemental Digital Content Table 1, available at: http://links.lww.com/JHTR/A362) indicate that multiple demographics, all mood and behavioral health variables, and a majority of the neuropsychological test scores independently predicted WHODAS 2.0 functional ratings. We also estimated correlation coefficients, which were consistent with the unadjusted regression findings (see Supplemental Digital Content Table 1, available at: http://links.lww.com/JHTR/A362). Notably, multiple indexes from the CPT-II (measuring reaction time and consistent attention) and WCST-64 Total Error scores (a measure of problem-solving/adaptive reasoning) were not significantly associated with WHODAS 2.0 function in these analyses.

In a series of multiple adjusted regression models, the observed relationships between neuropsychological test performance and WHODAS 2.0 functional ratings were much more limited (see Table 4). When each regression model included the neuropsychological test score as a covariate with the demographic, mood, and behavioral health covariates, verbal learning and memory (CVLT-II Total Trials 1–5 and CVLT-II Delayed Recall scores) remained significant predictors of function (see Table 4 for all adjusted regression results). More specifically, veterans with lower verbal learning and spontaneous delayed recall scores also reported lower overall functional abilities.

TABLE 4.

Significant adjusted individual predictors of WHODAS 2.0 regressions with the entire sample^a

	CVLT-II Trials 1–5			CVLT-II Delayed Recall
	Coefficient	95% CI	P	Coefficient	95% CI	P
Cognitive measure of interest	−0.190	−0.31 to −0.07	.001	−1.84	−2.88 to −0.80	.001b
Age	0.17	0.03 to 0.31	.020	0.17	0.03 to 0.31	.017c
Total no. symptoms (NSI)	1.11	0.85 to 1.37	.000	1.12	0.86 to 1.38	.000b
% income disability	3.63	1.38 to 5.88	.002	3.77	1.53 to 6.01	.001b
Depression	3.40	1.80 to 5.01	.000	3.41	1.81 to 5.01	.000b
Insomnia	3.85	1.16 to 6.54	.005	3.60	0.94 to 6.27	.008d

Abbreviations: CVLT-II, California Verbal Learning Test–Second Edition; FSIQ, full-scale IQ; NSI, Neurobehavioral Symptom Inventory; PTSD, posttraumatic stress disorder; TBI, traumatic brain injury; WAIS-III, Wechsler Adult Intelligence Scale–Third Edition; WHODAS, World Health Organization Disability Assessment Scale.

^aOnly significant findings reported; predicted WAIS-III FSIQ, sex, race, ethnicity, education, marital status, predeployment employment, current employment, anxiety, PTSD, positive TBI history, alcohol use, and pain were not significant. Cognitively, all neuropsychological measures aside from the 2 CVLT-II learning and memory variables were not significant.

^bP ≤ .001.

^cP ≤ .05.

^dP ≤ .01.

Predictors of WHODAS 2.0 functional disability ratings within the deployed control and clinical (mTBI/PTSD) subgroups

Similar regression models were conducted with subgroups. Predictors of functioning and disability for each of the 3 clinical groups and the deployed control group were evaluated. Significant findings from these models are summarized in Table 5, and all results are detailed in Supplemental Digital Content Table 2 (available at: http://links.lww.com/JHTR/A362).

TABLE 5.

Significant predictors of WHODAS 2.0 measure of functional disability within the clinical and control groups—Adjusted regressions^a

Variables	Deployed control	mTBI + PTSD	PTSD-only	mTBI-only
Cognitive variables of interest (predictor variables)	E, J, K	L
Age	A, C, G, J, K
NSI symptoms	A, B, C, D, E, F, G, H, I, J, K, L, M	A, B, C, D, E, F, G, H, I, J, K, L, M
Predeployment employment				J
Receiving disability	A, J, K, L, M
Depression	A, B, C, D, E, F, G, H, I, J, K, M	A, G, H, I, J, K, L, M	A, B, C, D, E, F, G, I, J, K, L, M
Anxiety		L

Abbreviations: CPT-II, Conners’ Continuous Performance Test–Second Edition; CVLT-II, California Verbal Learning Test–Second Edition; D-KEFS, Delis-Kaplan Executive Function System; FSIQ, full-scale IQ; mTBI, mild traumatic brain injury; NSI, Neurobehavioral Symptom Inventory; PTSD, posttraumatic stress disorder; WAIS-III, Wechsler Adult Intelligence Scale–Third Edition; WAIS-IV, Wechsler Adult Intelligence Scale–Fourth Edition; WCST, Wisconsin Card Sorting Test; WHODAS, World Health Organization Disability Assessment Scale.

^aEach letter (defined in the following key) represents the regression model for a specific neuropsychological test as the predictor variable, with WHODAS 2.0 function as the dependent variable. Each model included all demographic and behavioral health variables as covariates. Only significant findings are reported; WTAR predicted WAIS-III FSIQ, sex, race, ethnicity, education, marital status, current employment, alcohol use, insomnia, and pain were not significant across adjusted regressions for any clinical group; italicized = P ≤ .05; underlined = P ≤ .01; bold = P ≤ .001. Neuropsychological tests are coded as follows: (A) D-KEFS Trail Making: Number-LetterSwitching; (B) D-KEFS Verbal Fluency: Letter Fluency; (C) D-KEFS Verbal Fluency: Category Switching Total Correct; (D) D-KEFS Verbal Fluency: Category Switching Total Switching Accuracy; (E) D-KEFS Color-Word Interference Test: Inhibition; (F) D-KEFS Color-Word Interference Test: Inhibition/Switching; (G) Conners’ CPT-II: Omissions; (H) Conners’ CPT-II: Commissions; (I) Conners’ CPT-II: Hit Rate; (J) CVLT-II: Trials 1–5; (K) CVLT-II: Delayed Recall; (L) WCST: Total Errors; (M) WAIS-IV: Processing Speed Index.

In the results from the multiple regressions for clinical groups, there is limited evidence of relationships between cognitive capacities and WHODAS 2.0 functional abilities. For the mTBI + PTSD group, only one of 13 neuropsychological tests was associated with WHODAS 2.0 ratings. Specifically, better WCST-64 Total Error scores were positively associated with lower functioning. For this same group, higher levels of NSI symptoms (13/13 models), depression (8/13), and anxiety (1/13) predicted lower functioning/increased disability. In the PTSD-only and mTBI-only groups, none of the 13 neuropsychological test scores significantly predicted WHODAS 2.0 functional abilities when controlling for covariates. Within the PTSD-only group, levels of depression were significantly associated with functional abilities in 12 of 13 models. Within the mTBI-only group, predeployment employment status significantly predicted functional abilities in one of 13 models.

Consistent with results from examination of the total sample, the deployed control group’s performance on tests of verbal learning (CVLT-II Total Trials 1–5) and memory (CVLT-II Delayed Recall) was significantly associated with WHODAS 2.0 functional ratings, where better test performance predicted better everyday functional abilities. Performance on an executive functioning test of response inhibition also significantly predicted functional abilities, with decreased inhibition associated with decreasing functional abilities. Age (5/13 models), number of NSI symptoms (13/13), receipt of disability income (5/13), and depression symptoms (12/13) also predicted functional abilities.

DISCUSSION

The purpose of this study was to advance understanding of the predictors of functional abilities/disabilities in veteran populations, especially the role of cognitive functioning. To advance this understanding, we examined the relationships between WHODAS 2.0 scores, behavioral health and demographic covariates, and performance on a range of neuropsychological tests. Consistent with prior findings,⁴¹ unadjusted regression analyses showed significant associations between most neuropsychological test scores and WHODAS 2.0 functional disability ratings, suggesting a relationship between everyday functioning and capacities such as memory, processing speed, and executive functioning. In multiple regression models adjusted for mood, behavioral health, postconcussive symptoms, and demographics, however, the relationships between cognitive test performance and everyday functioning were drastically attenuated. Here, only 2 of the 13 neuropsychological test scores (verbal learning and memory) remained significant predictors of WHODAS 2.0 functional abilities.

The most striking finding was how little neuropsychological test scores predicted WHODAS 2.0 functional abilities when covariates were controlled for and participants were stratified into clinical groups (mTBI-only, PTSD-only, and mTBI + PTSD) and compared with a deployed control group. In these multiple regression models estimated with each neuropsychological test, 3 of the 13 neuropsychological test scores (verbal learning, verbal memory, and executive function/inhibition) were significantly associated with WHODAS 2.0 functional abilities for the deployed control group. Surprisingly, however, none of the 13 neuropsychological test scores were significant predictors of WHODAS 2.0 functional abilities in the mTBI-only and PTSD-only groups, and only 1 neuropsychological test score (WCST-64 Total Errors) was significantly associated with WHODAS 2.0 functional abilities in the mTBI + PTSD group. In addition, the direction of the relationship in our mTBI + PTSD group, between cognitive test performance (WCST-64 Total Errors) and functioning, was counterintuitive, as higher levels of performance on a test of reasoning and problem solving were associated with worsening WHODAS 2.0 functional abilities. This finding is difficult to interpret, but we speculate that stronger reasoning and problem-solving skills may be related to greater awareness and more accurate reporting of functional limitations.

Similar to our neuropsychological test score findings, significant relationships between many demographic, mood, behavioral health, and postconcussive symptoms (see Tables 1 and 2) and WHODAS 2.0 ratings were observed in unadjusted analyses. However, there was also a substantive reduction in the number of significant findings with multiple adjusted models in individual clinical groups. Conversely, in the deployed control group, many variables remained predictive of WHODAS 2.0 functional abilities in the models that controlled for veteran characteristics, including age, total NSI symptoms, receipt of disability monies, and depression, as well as neuropsychological tests scores for verbal learning, memory, and inhibition. The association between NSI postconcussive symptoms and everyday functioning was unexpected for the deployed control group, although prior research demonstrates appreciable rates of postconcussive symptoms in OIF veterans without mTBI or PTSD⁴² and underscores the fact that many common postconcussive symptoms are not necessarily specific to concussion (eg, headaches, irritability).

Prior research has not directly addressed the relationship between neuropsychological test performance and everyday functioning in OEF/OIF/OND veteran populations. However, subjective cognitive complaints and mixed evidence for measurable cognitive impairment are reported for veterans with mTBI and PTSD.²¹ More specifically, cognitive dysfunction is thought to contribute to postmilitary functional difficulties. Our findings provide important context to this literature, indicating that tested levels of various cognitive capacities may correlate with levels of functional disability but that, at least in the case of veterans with histories of mTBI and/or PTSD, other factors such as depression, postconcussive NSI symptoms, and anxiety are more strongly related to self-reported everyday functioning.

Consistent with prior research,⁴³ significant differences were observed across our clinical groups on most neuropsychological tests, with participants in the mTBI + PTSD group tending to obtain the lowest scores across tests. Importantly, even these clinical group mean scores across all 13 administered neuropsychological tests were still within “normal” performance ranges (ie, within 1 standard deviation of normative population data for the measure). This feature of our data seems well aligned with prior research,²¹ and our broader findings that objectively measured cognitive capacities are not primary determinants of functional limitations experienced by OIF/OED/OND veterans with mTBI and/or PTSD. While our data suggest that neuropsychological testing may be useful for establishing the extent to which subjective cognitive complaints are rooted in measurable deficits on objective testing, the discrepancies between objective test performance and self-report of functioning also suggest a need for further research with more ecologically valid measures of applied cognition and task performance.^44,45

In practical terms, our findings may also have implications for intervention planning. Although further study is necessary before firm recommendations can be made, our results suggest that relative to cognitive symptoms, other postconcussive and depression symptoms may represent promising treatment targets for OEF/OIF/OND veterans with mTBI and PTSD who are experiencing functional limitations. Furthermore, in line with previous work on patient education following mTBI,^46,47 the high rates of persisting postconcussive symptoms observed in clinical groups suggest value in early psychoeducational interventions regarding expectations for mTBI recovery, which have been shown to reduce rates of reported postconcussive symptoms. Depression and postconcussive symptoms may also represent relevant treatment targets.

This study included a number of limitations. First, the assessment of functioning was limited to self-report. Although WHODAS 2.0 is a valid self-report measure of daily functioning and disability in OIF/OEF/OND veteran populations,⁴¹ recent findings indicate that the test may not capture mild functional disability related to mTBI and PTSD. In addition, our study did not control for functional impairments due to physical injuries (eg, loss of limb/amputation, spinal cord injury), which may have impacted WHODAS 2.0 ratings and could reflect the reason for receiving disability payments. Other limitations include small sample sizes in the clinical subgroups, potential ecological validity issues with administered tests, variability in the time since deployment and injury, and limited generalizability due to the demographic characteristics of the sample.

CONCLUSION

Clarifying the factors that predict everyday functional abilities in OEF/OIF/OND veterans with and without mTBI and/or PTSD is an important task that has implications for the planning, design, and implementation of treatment and services. Consistent with prior research on functional outcomes in veterans, our study indicates that numerous factors are associated with self-reported levels of functional abilities. However, for veterans with mTBI and/or PTSD, neuropsychological test performance does not appear to be a strong predictor of functional abilities relative to other relevant clinical variables. While some cognitive test scores predicted WHODAS 2.0 ratings in our deployed control group, among participants with PTSD and/or mTBI, our results suggest that depression and noncognitive symptoms are likely to be more important therapeutic targets when providing interventions to enhance daily function. For future research, replication of our results with more objective assessment of functional abilities and comprehensive assessment of other potential contributing factors to disability is recommended. In addition, evaluation of depression and postconcussive treatment outcomes on everyday functioning in this population is needed.