JRRD Volume 51, Number 1, 2014

Pages 59–70

Cognitive Symptom Management and Rehabilitation Therapy

(CogSMART) for Veterans with traumatic brain injury: Pilot randomized
controlled trial

Elizabeth W. Twamley, PhD;1–2* Amy J. Jak, PhD;1–3 Dean C. Delis, PhD;1 Mark W. Bondi, PhD;2–3 James B.
Lohr, MD1–2
1Center of Excellence for Stress and Mental Health, Department of Veterans Affairs (VA) San Diego Healthcare System,
San Diego, CA; 2Department of Psychiatry, University of California, San Diego, San Diego, CA; 3Psychology Service, VA
San Diego Healthcare System, San Diego, CA

Abstract—Traumatic brain injury (TBI) can result in cognitive INTRODUCTION

impairments and persistent postconcussive symptoms that
limit functional recovery, including return to work. We evalu- Traumatic brain injury (TBI) is considered the signa-
ated a 12 wk compensatory cognitive training intervention
ture wound among Veterans of Operation Iraqi Freedom
(Cognitive Symptom Management and Rehabilitation Therapy
[CogSMART]) in the context of supported employment for
(OIF) and Operation Enduring Freedom (OEF), mainly
Veterans with mild to moderate TBI. Participants were randomly because of the large numbers of blast injuries caused by
assigned to receive 12 wk of supported employment plus improvised explosive devices [1]. The vast majority of
CogSMART or enhanced supported employment that controlled these TBIs (89%) are in the mild or moderate range [1],
for therapist attention (control). CogSMART sessions were deliv- with estimates that approximately 20 percent of OIF/OEF
ered by the employment specialist and included psychoeducation
regarding TBI; strategies to improve sleep, fatigue, headaches,
and tension; and compensatory cognitive strategies in the
domains of prospective memory, attention, learning and memory,
and executive functioning. Compared with controls, those Abbreviations: CAPS = Clinician-Administered Posttraumatic
assigned to supported employment plus CogSMART demon- Stress Disorder Scale, CogSMART = Cognitive Symptom
strated significant reductions in postconcussive symptoms Management and Rehabilitation Therapy, CVLT-II = Califor-
(Cohen d = 0.97) and improvements in prospective memory func- nia Verbal Learning Test-2nd edition, D-KEFS = Delis-Kaplan
tioning (Cohen d = 0.72). Effect sizes favoring CogSMART for Executive Function System, DOD = Department of Defense,
posttraumatic stress disorder symptom severity, depressive symp- HAM-D = Hamilton Depression Rating Scale, IQ = intelligence
tom severity, and attainment of competitive work within 14 wk quotient, IRB = Institutional Review Board, LOC = loss of
were in the small to medium range (Cohen d = 0.35–0.49). Those consciousness, MIST = Memory for Intentions Screening Test,
who received CogSMART rated the intervention highly. Results NSI = Neurobehavioral Symptom Inventory, OEF = Operation
suggest that adding CogSMART to supported employment may Enduring Freedom, OIF = Operation Iraqi Freedom, PTSD =
improve postconcussive symptoms and prospective memory. posttraumatic stress disorder, TBI = traumatic brain injury, VA =
These effects, as well as smaller effects on psychiatric symptoms Department of Veterans Affairs.
and ability to return to work, warrant replication in a larger trial. *
Address all correspondence to Elizabeth W. Twamley,
PhD; Center of Excellence for Stress and Mental Health,
Key words: attention, brain injury, cognitive rehabilitation, cog- VA San Diego Healthcare System, 3350 La Jolla Village Dr
nitive training, depression, employment, executive functioning, (116A), San Diego, CA 92161; 858-642-3848; fax: 858-642-
memory, postconcussive symptoms, posttraumatic stress disor- 1243. Email: etwamley@[Link]
der, rehabilitation, unemployment. [Link]

59
60

JRRD, Volume 51, Number 1, 2014

Veterans have sustained a mild TBI [2–3]. Numerous remember activities). Although there have been numerous
studies demonstrate that individuals with mild to moder- studies of cognitive rehabilitation in TBI [24–25], most
ate TBI exhibit problems with concentration, learning published TBI rehabilitation research has involved people
and memory, prospective memory (remembering to do with severe TBI, and little is known about rehabilitation
things in the future), and problem solving [4–8], all of for those with mild to moderate TBI [26]. There has been
which can limit functional recovery, including cognitive only one randomized controlled trial of comprehensive
readiness for work and school [9–10]. However, there has cognitive rehabilitation therapy in a group with mild to
been scant research on cognitive rehabilitation for indi- moderate TBI with chronic symptoms (i.e., patients whose
viduals with mild to moderate TBI. injuries occurred more than 6 mo before treatment [27]),
Studies show that neurocognitive functioning is which is the population most similar to OIF/OEF Veterans
strongly associated with employment status across a vari- with TBI histories presenting for care at VA facilities. As a
ety of disorders (e.g., human immunodeficiency virus, result of the limited high-quality research available, the
epilepsy, and psychiatric disorders), including TBI, with VA/Department of Defense (DOD) Clinical Practice
effect sizes in the medium range (executive functioning: Guideline for Management of Concussion/Mild TBI [28]
d = 0.62, learning and memory: d = 0.61, and attention/ does not make recommendations in favor of or against
concentration: d = 0.53 [11]). Difficulty learning job compensatory training or training on the use of external
tasks, distractibility, or slowness in job performance, for memory aids; it does, however, emphasize psychoeduca-
example, can lead to job failures, and cognitive abilities tion and family education regarding TBI, as well as func-
among those with TBI histories have been shown to pre- tional and vocational interventions to promote community
dict functional skills and ability to work [12–14]. reintegration [28].
In addition, recent data show that about 7 percent of Based on our reviews of prior cognitive training stud-
OIF/OEF Veterans who use Department of Veterans ies [19–20,22], we developed a manualized, 12 wk, multi-
Affairs (VA) healthcare have persistent postconcussive modal compensatory cognitive training intervention
symptoms such as difficulties with cognitive functioning, emphasizing habit learning and compensatory strategies
sleep, fatigue, emotional functioning, and headaches [15]. in prospective memory, attention, learning and memory,
Healthcare costs and service utilization rates are greater in and executive functioning. The treatment manual was
these Veterans, possibly because of high rates of comor- informed by consultation with the acquired brain injury
bidities such as posttraumatic stress disorder (PTSD) program at Mesa College in San Diego, California, and
(comorbid in 73%); depression (47%); and back, neck, other cognitive remediation experts. The compensatory
and/or headache pain (72%) [15]. These comorbidities cognitive training intervention has demonstrated efficacy
can also further contribute to cognitive impairment [4,16– in improving cognition, psychiatric symptoms, functional
17]. Thus, a need exists for interventions to improve cog- capacity, and quality of life in people with severe mental
nition and functioning of these individuals. There are no illness [29]. The compensatory cognitive training manual
known medications that correct the cognitive deficits of was subsequently adapted for Veterans with mild to mod-
TBI. Given the extraordinary plasticity of the human brain erate TBI; consistent with the Clinical Practice Guideline
[18], cognition must be considered one of the most prom- [28], sections on psychoeducation regarding TBI and
ising targets for improvement via psychological methods. strategies to improve postconcussive symptoms (sleep
Cognitive training is a rehabilitative technique that has disturbance, fatigue, headaches, and tension) were added,
been used in TBI, psychiatric disorders [19–21], and mem- and the entire manual was rewritten to be applicable to
ory disorders such as Alzheimer disease [22]. The research Veterans with TBI. Prospective memory, or memory for
on cognitive training interventions has examined both intentions, continued to be highlighted because it is one of
restorative and compensatory approaches to treatment the most frequent sequelae of TBI and because of its obvi-
[23]. Restorative interventions aim to restore cognitive ous implications for treatment adherence [30]. The result-
abilities via drills and practice, whereas compensatory ing intervention, Cognitive Symptom Management and
interventions teach clients strategies to work around their Rehabilitation Therapy (CogSMART), is portable, practi-
cognitive deficits, using alternative ways to carry out daily cal, and designed to be implemented without extensive
living activities. Compensatory strategies may be internal training. CogSMART can be delivered to individuals or
(e.g., using acronyms to remember information) or exter- groups, in the clinic or in the community, increasing the
nal (e.g., using a calendar, alarms, or smartphones to potential for training gains to transfer into the real-world
 61

TWAMLEY et al. CogSMART for Veterans with TBI

environment. CogSMART encourages the involvement of may improve functional outcomes because compensatory
family or other support persons in the first two and last interventions can be individualized and tailored to each
two sessions of the intervention, so that family members person’s job search process and job duties.
can learn about TBI, generate a plan for supporting the In our pilot study, all participants received supported
patient’s use of new strategies, give feedback about the employment for 1 yr. During the first 3 mo of the study,
effectiveness of strategy use, and plan for further applica- participants were randomized to also receive CogSMART
tion of the strategies to daily life activities. Table 1 lists or additional supported employment sessions (enhanced
the CogSMART modules and examples of strategies supported employment). We hypothesized that, compared
taught. with enhanced supported employment, supported employ-
Meta-analytic studies of cognitive rehabilitation for ment plus CogSMART would result in reductions in post-
severe mental illness have shown that cognitive interven- concussive symptoms (primary outcome), improvements
tions lead to better outcomes when they are embedded in a in cognition and functional capacity, and better rates of
broader psychosocial rehabilitation program [20–21], and
job placement at postintervention. In this article, we pres-
the Clinical Practice Guideline recommended vocational
ent the baseline and postintervention data; future reports
interventions to improve community integration [28]. Fur-
will address the durability of CogSMART effects and its
thermore, compensatory strategies have been used suc-
longer-term effects on work outcomes over 1 yr.
cessfully within the context of supported employment for
individuals with TBI [31]. Thus, we chose to evaluate the
efficacy of CogSMART within the context of an evidence-
based supported employment program. Supported employ- METHODS
ment is a cost-effective, evidence-based, individualized
approach to work rehabilitation emphasizing rapid job Participants
searching for competitive work in the community, based Fifty Veterans receiving healthcare at the VA San
on client interests and preferences [9,32]. Augmenting Diego Healthcare System enrolled in the study and gave
work rehabilitation with compensatory cognitive training their written informed consent prior to study participation.

Table 1.
Domains targeted in Cognitive Symptom Management and Rehabilitation Therapy (CogSMART) and their associated strategies.
Impaired Domain Specific Compensatory Strategies and Habits Taught in CogSMART
Postconcussive Symptoms Psychoeducation regarding natural course of postconcussive symptoms.
Appropriate pacing, use of routines, and lifestyle strategies.
Stress reduction (e.g., progressive muscle relaxation, abdominal breathing, mindfulness, visualization,
grounding).
Sleep hygiene education; headache management; and education regarding depression, anxiety, and
posttraumatic stress disorder.
Prospective Memory Daily calendar use.
To-do lists and prioritizing tasks.
Linking tasks and using “can’t miss reminders” to cue tasks.
Attention and Vigilance Conversational vigilance skills (reduce distractions, eye contact, paraphrasing, and asking questions).
Task vigilance skills (paraphrase instructions, use self-talk during tasks to maintain focus).
Learning and Memory Encoding strategies (write things down, paraphrasing/repetition, association, chunking, categorizing,
acronyms, rhymes, visual imagery, name-learning strategies).
Retrieval strategies (systematic searching) and organizational strategies for general learning and
memory.
Executive Functioning 6-step problem-solving method (define problem, brainstorm solutions, evaluate solutions, select
solution, try it, evaluate how it worked).
Self-talk while solving problems.
Hypothesis testing and self-monitoring.
62

JRRD, Volume 51, Number 1, 2014

Inclusion criteria were (1) OIF/OEF Veteran; (2) history Procedure

of mild to moderate TBI (loss of consciousness [LOC] Study referrals came from the VA San Diego Health-
<6 h; posttraumatic amnesia <7 d) according to the Clini- care System Wellness and Vocational Enrichment Clinic,
cal Practice Guideline [28], documented in a prior clinical TBI Cognitive Rehabilitation Clinic, Polytrauma Clinic,
neuropsychological evaluation and confirmed by a struc- and Neuropsychological Assessment Unit. Participants
tured interview; (3) documented impairment (>1 standard were compensated $20 per assessment session but were
deviation below the mean) in at least one neuropsycho- not paid to participate in treatment.
logical domain (i.e., attention, processing speed, working Following baseline assessment, participants were ran-
memory, learning, memory, executive functioning), as domized to one of two conditions: supported employment
determined by valid clinical neuropsychological testing plus CogSMART or enhanced supported employment.
by a VA or DOD neuropsychologist using at least one Randomization was carried out by the principal investigator
effort test (e.g., Test of Memory Malingering, California using a randomization scheme generated by Randomiza-
Verbal Learning Test-2nd edition [CVLT-II] Forced [Link], with 50 participants in one block. All Veterans in
Choice); and (4) unemployed, but stating a goal of work. the study received supported employment for 1 yr, the goal
Veterans who met criteria for current alcohol and/or sub- of which is competitive employment. Two supported
stance abuse or dependence or who were participating in employment specialists provided all services; one employ-
other intervention studies were excluded. Eight partici- ment specialist delivered CogSMART for 1 h/wk in addi-
pants dropped out, four from each group (two decided not tion to standard supported employment (i.e., 2 visits/wk),
to pursue work, one moved, and five were lost to follow- and the other employment specialist delivered enhanced
up). Posttreatment data were available for 34 participants supported employment (2 visits/wk) to control for the non-
at 3 mo (16 in supported employment plus CogSMART [3 specific therapeutic factors provided in CogSMART.
with moderate TBI] and 18 in enhanced supported CogSMART and enhanced supported employment were
employment [4 with moderate TBI]). The 34 participants provided during the first 12 wk of supported employment
with complete baseline and posttreatment data did not dif- so that time and contact with the employment specialist
fer on age, education level, sex, race, ethnicity, premorbid were equivalent across groups. Consistent with the sup-
intelligence quotient (IQ), length of LOC during their ported employment model, services were offered at loca-
worst TBI, or length of total LOC summed across up tions of the participant’s choosing (e.g., career center,
to four TBIs (all p > 0.13) from the 16 participants who home, coffee shop, library, or VA clinic). All participants
did not have a 3 mo assessment. On average, the received standard clinical care with their usual providers
34 participants with complete data were 32 yr old, had during the trial. CogSMART completion rates were high;
13.6 yr of education, were 94 percent male, and were 15 of the 16 participants randomized to receive supported
76 percent members of a racial or ethnic minority group; employment plus CogSMART completed all 12 sessions,
76 percent met criteria for threshold PTSD [33]. The and 1 participant completed 8 sessions.
median length of LOC for the worst TBI was 1.5 min, and To ensure CogSMART treatment fidelity, all
82 percent of these injuries were contact TBIs (vs blast CogSMART sessions were audiotaped and 20 percent were
only); the median length of total LOCs summed across up randomly selected for fidelity rating every 2 wk. Adherence
to four TBIs was 1.7 min. Table 2 presents sample charac- rates were consistently in the 90 to 100 percent range.
teristics by group. The treatment groups did not differ by
sex, race, ethnicity, postconcussive and psychiatric symp- Measures
tom severity, presence of mild or greater depressive symp- Measures were administered at baseline and 3 mo (i.e.,
toms, presence of threshold PTSD, TBI severity (length of following completion of the CogSMART or enhanced sup-
LOC in the worst TBI and summed across up to 4 TBIs), ported employment portion of the study). All tests were
years since their most recent TBI, nature of their worst administered according to standardized procedures by a
TBI (contact vs blast only), or years since their worst TBI research assistant trained to a high level of interrater reli-
(all p  0.06). The group that received supported employ- ability (i.e., >0.90). The research assistant was not aware
ment plus CogSMART, however, was about 5 yr younger of participant randomization status at the baseline assess-
on average than the group that received enhanced sup- ment but was aware of treatment group for subsequent
ported employment (p = 0.05). assessments.
 63

TWAMLEY et al. CogSMART for Veterans with TBI

Table 2.
Sample characteristics and differences between study groups at baseline assessment.
t-Value,
Supported Employment + Enhanced Supported
Characteristic Mann-Whitney df p-Value
CogSMART (n = 16) Employment (n = 18)
U, or χ2
Age, yr (mean ± SD) 29.4 ± 6.2 34.3 ± 7.4 2.1 32 0.05
Education, yr (mean ± SD) 13.0 ± 1.4 14.1 ± 1.8 1.9 32 0.06
Sex: Male (%) 93.8 94.4 <0.1 1 0.93
Race: Caucasian (%) 56.3 61.1 6.1 3 0.11
Ethnicity: Hispanic/Latino (%) 37.5 33.3 0.1 1 0.80
Length of LOC of Worst TBI, min (median [range]) 6.5 (0–420) 1.0 (0–1,440) 124.5 — 0.51
Length of Total LOC Across 4 TBI, min (median [range]) 6.5 (0–420) 1.0 (0–1,440) 120.0 — 0.42
Time Since Worst TBI, yr (mean ± SD [range]) 4.4 ± 3.3 (0–14) 6.3 ± 5.6 (1–19) 1.2 28* 0.23
Time Since Most Recent TBI, yr (mean ± SD [range]) 3.6 ± 2.7 (0–9) 5.1 ± 5.3 (1–19) 1.1 32 0.29
NSI (mean ± SD) 41.7 ± 16.8 34.2 ± 11.4 1.5 25.9* 0.14
CAPS (mean ± SD) 64.3 ± 23.0 53.4 ± 26.2 1.3 32 0.21
HAM-D (mean ± SD) 17.3 ± 5.4 13.4 ± 6.8 1.8 32 0.08
Premorbid IQ Estimate (WRAT-3 Reading) (mean ± SD) 95.1 ± 7.8 96.5 ± 11.3 0.4 32 0.67
MIST Summary Score (mean ± SD) 33.6 ± 10.3 32.7 ± 9.3 0.3 32 0.80
MIST 24-Hour Probe (mean ± SD) 0.3 ± 0.6 0.6 ± 0.9 1.1 29.6* 0.27
WAIS-III Digit Span Scaled Score (mean ± SD) 8.5 ± 2.1 9.7 ± 3.4 1.1 31 0.27
CVLT-II Trials 1–5 Learning T-Score (mean ± SD) 42.4 ± 11.3 40.2 ± 9.3 0.6 32 0.54
CVLT-II Long Delay Free Recall Z-Score (mean ± SD) 1.2 ± 1.3 1.1 ± 1.0 <0.1 32 0.97
D-KEFS Letter Fluency Scaled Score (mean ± SD) 9.3 ± 2.5 8.8 ± 3.1 0.5 32 0.63
D-KEFS Category Fluency Scaled Score (mean ± SD) 8.5 ± 2.3 8.0 ± 4.2 0.4 27* 0.67
D-KEFS Category Switching Scaled Score (mean ± SD) 8.0 ± 3.2 6.8 ± 2.3 1.2 32 0.23
WCST-64 Perseverative Errors T-Score (mean ± SD) 45.9 ± 6.4 42.2 ± 10.8 1.3 28.2* 0.22
*Adjusted to unequal variances between groups.
CAPS = Clinician-Administered Posttraumatic Stress Disorder Scale, CogSMART = Cognitive Symptom Management and Rehabilitation Therapy, CVLT-II = Cali-
fornia Verbal Learning Test-2nd edition, df = degrees of freedom, D-KEFS = Delis-Kaplan Executive Function System, HAM-D = Hamilton Depression Rating Scale,
IQ = intelligence quotient, LOC = loss of consciousness, MIST = Memory for Intentions Screening Test, NSI = Neurobehavioral Symptom Inventory, SD = standard
deviation, TBI = traumatic brain injury, WAIS-III = Wechsler Adult Intelligence Scale-3rd edition, WCST-64 = Wisconsin Card Sorting Test-64 card version, WRAT-
3 = Wide Range Achievement Test-3rd edition.

Measures of cognitive functioning were selected to spective memory ability during the 30 min test, and
appropriately characterize the sample and assess change in the 24 h raw probe score reflects long-term prospec-
the four cognitive domains targeted by the CogSMART tive memory ability over a 24 h period (the examinee
intervention (prospective memory, attention, learning and is asked to call the examiner 24 h following the test to
memory, and executive functioning), as follows: report how he or she slept).
1. Premorbid IQ was estimated with the Wide Range 3. Attention and working memory were measured with
Achievement Test-3rd edition (WRAT-3) [34] Read- the Wechsler Adult Intelligence Scale-3rd Edition
ing test, which requires the examinee to correctly
[36] Digit Span scaled score, which requires the
pronounce irregularly spelled English words and was
examinee to repeat strings of digits in both forward
administered at baseline only because this ability was
not expected to change over time. The age-corrected and backward order.
standard score is reported. 4. Verbal learning and memory were measured with the
2. Prospective memory was measured with the Memory CVLT-II [37]. The trials 1–5 T-score was used as an
for Intentions Screening Test (MIST) [35], which index of verbal learning and reflects acquisition of
requires examinees to perform directed actions at cer- words over five learning trials. The long delay free
tain times or in response to specific cues. The MIST recall Z-score was used to measure memory follow-
raw summary score is an index of short-term pro- ing a 20 min delay period.
64

JRRD, Volume 51, Number 1, 2014

5. Executive functioning was measured with the Delis- ables were not normally distributed, so medians were
Kaplan Executive Function System (D-KEFS) [38] used to characterize the sample. Differences between
Verbal Fluency test, which requires examinees to groups were analyzed with t-tests, Mann-Whitney U
generate words beginning with a given letter (letter tests, and chi square tests. Although the treatment groups
fluency) or belonging to a given category (category differed in age (Table 2), we did not adjust for this differ-
fluency) and to switch between words from two dif- ence in our models because of the pilot nature of the
ferent categories (category switching); age-corrected study and the small sample size. Additionally, the age
scaled scores were used for these measures. The Wis- range in the two groups was similar (22–49 in the
consin Card Sorting Test-64 card version [39] was enhanced supported employment group and 22–44 in the
also used to measure reasoning and set-switching; the supported employment plus CogSMART group). The
perseverative errors T-score correcting for age and outcome data were analyzed using t-tests between mean
education was used. change scores in each group and chi square tests for work
Postconcussive symptom severity was measured with outcomes (attainment of competitive employment within
the Neurobehavioral Symptom Inventory (NSI) (a VA 14 wk, which represented the average amount of time
instrument developed following Cicerone and Kalmar between the baseline and postintervention assessment).
[17]), which measures severity of self-reported cognitive, Secondary analyses removing the two oldest individuals
physical, and emotional symptoms on a 0 to 4 scale ranging in the enhanced supported employment group (which
from “none” to “very severe.” There are 12 physical symp- made the two groups not significantly different on age)
toms (e.g., dizziness, headaches, nausea, light and sound did not change the results.
sensitivity), 4 cognitive symptoms (e.g., problems with
concentration, memory, decision-making, slowed thinking),
and 6 emotional symptoms (e.g., fatigue, sleep disturbance, RESULTS
anxiety, depression, irritability). PTSD symptom severity
was measured with the Clinician-Administered PTSD Independent samples t-tests on change scores between
Scale (CAPS) [40], the gold-standard interview to deter- groups revealed significant CogSMART-associated
mine PTSD diagnosis and measure symptom severity. The improvements in postconcussive symptoms (NSI: p =
CAPS total score symptom severity descriptors are 0–19 = 0.01) and prospective memory performance (MIST 24 h
asymptomatic, 20–39 = mild/subthreshold, 40–59 = moder- probe: p = 0.05) at posttreatment (Table 3). The Cohen d
ate/threshold, 60–79 = severe, and 80 = extreme [33]. effect sizes for these group differences between change
Depressive symptom severity was measured with the 17- scores were 0.97 and 0.72, respectively. The remainder of
item Hamilton Depression Rating Scale (HAM-D) [41]. the neuropsychological, symptom severity, quality of life,
The HAM-D total score symptom severity descriptors are and work outcome comparisons did not reveal statistically
0–7 = normal, 8–13 = mild, 14–18 = moderate, 19–22 = significant differences between groups. However, the sup-
severe, and 23 = very severe [41]. The Quality of Life ported employment plus CogSMART group showed small
Interview-Brief Version [42] assessed subjective judgment to medium effect size improvements in psychiatric symp-
of global quality of life, rated by the participants on a 1 to tom severity (CAPS: d = 0.43 and HAM-D: d = 0.37,
7 scale ranging from “terrible” to “delighted.” Data regard- based on group differences between change scores) rela-
ing job attainment, hours worked, and wages earned were tive to the enhanced supported employment group. Five
collected weekly. Participants randomized to the supported participants in the enhanced supported employment con-
employment plus CogSMART condition rated the compo- dition obtained competitive work within the first 14 wk
nents of the CogSMART intervention on a scale from 1 to 5 of the study compared with eight participants in the sup-
(“not helpful” to “extremely helpful”) and provided written ported employment plus CogSMART condition (d =
qualitative comments following completion of the 0.49) (Table 3).
CogSMART sessions. Those who received CogSMART rated it highly with
regard to helpfulness of information regarding TBI and
Analyses postconcussive symptoms, information on PTSD, head-
Prior to the analyses, data were examined for missing ache strategies, fatigue strategies, sleep strategies, pro-
values, statistical outliers, and normality. The LOC vari- spective memory strategies, attention strategies, learning
 65

TWAMLEY et al. CogSMART for Veterans with TBI

Table 3.
Comparisons of change in scores (posttreatment–baseline) and competitive work attainment between groups.
Supported Enhanced
Employment + Supported t-Value
Outcome Measure (mean ± SD) df p-Value Cohen d
CogSMART Employment or χ2
(n = 16) (n = 18)
MIST Summary Score 4.9 ± 7.1 5.4 ± 7.6 0.2 31 0.83 0.08
MIST 24-Hour Probe 0.1 ± 0.7 0.5 ± 0.9 2.0 31 0.05 0.72
WAIS-III Digit Span Scaled Score 0.0 ± 1.5 0.7 ± 1.6 1.2 31 0.23 0.43
CVLT-II Trials 1–5 Learning T-Score 6.9 ± 10.8 7.7 ± 9.1 0.2 31 0.82 0.08
CVLT-II Long Delay Free Recall Z-Score 0.7 ± 1.1 0.8 ± 1.0 0.2 31 0.89 0.05
D-KEFS Letter Fluency Scaled Score 0.6 ± 3.4 0.2 ± 2.4 0.8 31 0.42 0.28
D-KEFS Category Fluency Scaled Score 0.3 ± 3.7 0.6 ± 3.0 0.8 31 0.42 0.28
D-KEFS Category Switching Scaled Score 1.3 ± 4.7 0.7 ± 2.6 0.5 31 0.64 0.16
WCST-64 Perseverative Errors T-Score 0.1 ± 4.9 1.9 ± 8.0 0.9 31 0.40 0.30
NSI 7.9 ± 5.2 0.4 ± 9.6 2.7 31 0.01 0.97
CAPS 10.0 ± 17.8 2.4 ± 17.4 1.2 31 0.22 0.43
HAM-D 2.2 ± 4.5 0.6 ± 3.9 1.0 29 0.31 0.37
QOLI-Brief 0.3 ± 1.0 0.1 ± 1.0 0.6 31 0.55 0.21
Attained Competitive Work Within 14 wk (%) 50 26 2.1 1 0.15 0.49
Note: Cohen d values are positive if supported employment plus CogSMART group improved more than did enhanced supported employment group.
CAPS = Clinician-Administered Posttraumatic Stress Disorder Scale, CogSMART = Cognitive Symptom Management and Rehabilitation Therapy, CVLT-II = Cali-
fornia Verbal Learning Test-2nd edition, df = degrees of freedom, D-KEFS = Delis-Kaplan Executive Function System, HAM-D = Hamilton Depression Rating
Scale, MIST = Memory for Intentions Screening Test, NSI = Neurobehavioral Symptom Inventory, QOLI-Brief = Quality of Life Interview-Brief Version, SD = stan-
dard deviation, WAIS-III = Wechsler Adult Intelligence Scale-3rd edition, WCST-64 = Wisconsin Card Sorting Test-64 card version.

and memory strategies, problem-solving strategies, and All CogSMART participants indicated that they would
information regarding additional VA services (all means recommend CogSMART to other Veterans with similar
3.5 on a 5-point scale, where 3 = “moderately helpful” problems.
and 5 = “extremely helpful”). Prospective memory strate-
gies received the highest mean rating (4.3), and many
participants noted in their qualitative comments that they DISCUSSION
had begun using a calendar system (paper or smart-
phone). Sample comments from participants included— We found that CogSMART, in the context of supported
employment for Veterans with mild to moderate TBI, was
• “I wish CogSMART was utilized while we were still
associated with significant reductions in self-reported post-
in the service and that everyone had to go through it.”
concussive symptoms and improvements in real-world pro-
• “It has taken a lot of pressure off my wife because she spective memory performance (the ability to carry out an
doesn’t have to keep up with my schedule and it has assigned task 24 h later). There were also nonsignificant
given me confidence to go to school.” CogSMART-associated reductions in PTSD and depressive
• “It helped me relax, concentrate, and remember impor- symptoms and return to competitive employment within
tant things.” 14 wk (all of which demonstrated small to medium effect
sizes over 0.35). CogSMART participants rated the inter-
• “[The strategies helped me with] fixing my credit, stay-
vention as helpful and stated universally that they would
ing on top of class work, and applying for competitive
recommend it to other Veterans. The high rate of comorbid
jobs.” PTSD (76%) in our sample is not atypical in OIF/OEF Vet-
• “It helped me reduce stress greatly by having better erans with mild TBI [43]. The etiology of cognitive impair-
organizational skills; I get better rest and it helps me ments in those with mild TBI and comorbid PTSD and/or
concentrate.” depression appears mainly attributable to psychiatric
66

JRRD, Volume 51, Number 1, 2014

comorbidity [44], and cognitive training approaches for CogSMART-associated effects emerge later in the course
those with both TBI and PTSD merit further research [43]. of the study. We will also examine job acquisition, job
We want to emphasize that although CogSMART may tenure (weeks worked), and wages earned over the course
have some modest effects on symptoms of PTSD and of the study. In a separate study, we are also examining
depression, these effects may overlap with those on post- compensatory cognitive training as a group treatment for
concussive symptoms, and CogSMART is not a treatment Veterans with TBI.
for PTSD or depression.
The strengths of our study include its randomized
design; robust control group; and use of both subjective, CONCLUSIONS
patient-centered outcome measures (e.g., self-reported
postconcussive symptoms, psychiatric symptoms, and CogSMART is a 12 wk intervention to improve post-
quality of life) and objective outcome measures (e.g., neu- concussive symptoms (e.g., sleep disturbance, fatigue,
ropsychological tests and competitive work outcomes). headaches, and tension) and cognition in the domains of
There were also several limitations to our study. First, this prospective memory, attention, learning and memory, and
was a small pilot study that was affected by 16 percent executive functioning. At this point, we tentatively con-
dropout within the first 3 mo as well as missing data at the clude that CogSMART may improve postconcussive
3 mo assessment. Future analyses will use hierarchical lin- symptoms and prospective memory performance. Psy-
ear modeling to fully take advantage of all data over the choeducation regarding TBI and postconcussive symp-
four assessments over the course of the study. The groups toms and training in compensatory strategies appear to be
differed in age at baseline, which we elected not to control perceived by Veterans as helpful. Results of this pilot
for in our analyses due to the small sample size and similar study are promising, but the CogSMART intervention
age range in both groups. However, the two groups did not needs to be studied in a larger trial.
differ on other symptom or injury characteristics, such as
length of LOC or severity of postconcussive symptoms,
cognitive impairment, or comorbid psychiatric symptoms. ACKNOWLEDGMENTS
Outcome assessment was not blinded; however, most of
our outcome measures were either objective (neuropsy- Author Contributions:
chological test performance, attainment of competitive Study design: E. W. Twamley.
work) or reported by the participant, rather than rated by Data analysis: E. W. Twamley.
Data interpretation: E. W. Twamley, A. J. Jak, D. C. Delis, M. W. Bondi,
the examiner. Therapist factors were a potential confound. J. B. Lohr.
We considered having a separate “cognitive specialist” Drafting of treatment manual: A. J. Jak.
deliver the CogSMART intervention instead of the Drafting of manuscript: E. W. Twamley.
employment specialist, but we believed the CogSMART Critical revision of manuscript for important intellectual content:
intervention would be more efficacious in the context of E. W. Twamley, A. J. Jak, D. C. Delis, M. W. Bondi, J. B. Lohr.
supported employment as well as more cost-effective if Financial Disclosures: Dr. Delis receives royalties from the sale of
the CVLT-II and D-KEFS.
delivered by one provider, which was necessarily the Funding/Support: This material was based on work supported by the
employment specialist. An advantage to this approach is DOD (award W81XWH-08–2-0193).
that the employment specialist can continue to use and Additional Contributions: The authors acknowledge the assistance
reinforce CogSMART principles throughout supported of Drs. Dewleen G. Baker and Rebecca E. Williams for their support
employment. We considered having the two employment of the study, the helpful consultation of Jennifer Vasterling and Paul
Wehman, and the administrative support of Kelsey Thomas and
specialists each provide services to participants in each
Amber Gregory. They also thank Laurie Arnold, MS, CRC, for her
study condition, but we opted to have one employment work with our Veterans and recognize the previous contributions of
specialist affiliated with each treatment condition to pre- Sarah K. Noonan, Gauri N. Savla, and Dawn Schiehser to the
vent treatment contamination. Finally, our results may not CogSMART treatment manual.
generalize to individuals with severe TBI, people with TBI Institutional Review: This study was approved by the VA Rehabilita-
who do not want to work, or non-Veterans. tion Research and Development Institutional Review Board (IRB).
Participant Follow-Up: The authors have no plans to notify the study
Future analyses will address the durability of participants of the publication of this article because there were no plans
CogSMART’s effects over the course of the entire 12 mo for participant re-contact following study participation, thus re-contact
study and will provide answers regarding whether was not approved by the IRB.
 67

TWAMLEY et al. CogSMART for Veterans with TBI

REFERENCES 11. Kalechstein AD, Newton TF, van Gorp WG. Neurocogni-
tive functioning is associated with employment status: A
1. Fischer H; Library of Congress, Congressional Research Ser- quantitative review. J Clin Exp Neuropsychol. 2003;25(8):
vice. United States military casualty statistics: Operation Iraqi 1186–91. [PMID:14566590]
Freedom and Operation Enduring Freedom. Washington [Link]
(DC): Congressional Research Service, Library of Congress; 12. Leahy BJ, Lam CS. Neuropsychological testing and func-
2007. tional outcome for individuals with traumatic brain injury.
2. Hoge CW, McGurk D, Thomas JL, Cox AL, Engel CC, Brain Inj. 1998;12(12):1025–35. [PMID:9876862]
Castro CA. Mild traumatic brain injury in U.S. Soldiers [Link]
returning from Iraq. N Engl J Med. 2008;358(5):453–63. 13. Drake AI, Gray N, Yoder S, Pramuka M, Llewellyn M.
[PMID:18234750] Factors predicting return to work following mild traumatic
[Link] brain injury: A discriminant analysis. J Head Trauma Reha-
3. Tanielian TL, Jaycox L; RAND Corporation. Invisible bil. 2000;15(5):1103–12. [PMID:10970931]
wounds of war: Psychological and cognitive injuries, their [Link]
consequences, and services to assist recovery. Santa Monica 14. Andelic N, Stevens LF, Sigurdardottir S, Arango-Lasprilla
(CA): RAND Center for Military Health Policy Research; JC, Roe C. Associations between disability and employ-
2008. ment 1 year after traumatic brain injury in a working age
4. Binder LM. Persisting symptoms after mild head injury: A population. Brain Inj. 2012;26(3):261–69.
review of the postconcussive syndrome. J Clin Exp Neuro- [PMID:22372413]
psychol. 1986;8(4):323–46. [PMID:3091631] [Link]
[Link] 15. Taylor BC, Hagel EM, Carlson KF, Cifu DX, Cutting A,
5. Kinsella G, Murtagh D, Landry A, Homfray K, Hammond Bidelspach DE, Sayer NA. Prevalence and costs of co-
M, O’Beirne L, Dwyer L, Lamont M, Ponsford J. Everyday occurring traumatic brain injury with and without psychiat-
memory following traumatic brain injury. Brain Inj. 1996; ric disturbance and pain among Afghanistan and Iraq War
10(7):499–507. [PMID:8806010] Veteran V.A. users. Med Care. 2012;50(4):342–46.
[Link] [PMID:22228249]
6. Vanderploeg RD, Crowell TA, Curtiss G. Verbal learning [Link]
and memory deficits in traumatic brain injury: encoding, 16. Mittenberg W, DiGiulio DV, Perrin S, Bass AE. Symptoms
consolidation, and retrieval. J Clin Exp Neuropsychol. following mild head injury: Expectation as aetiology. J Neu-
2001;23(2):185–95. [PMID:11309672] rol Neurosurg Psychiatry. 1992;55(3):200–204.
[Link] [PMID:1564481]
7. Vanderploeg RD, Curtiss G, Belanger HG. Long-term neuro- [Link]
psychological outcomes following mild traumatic brain 17. Cicerone KD, Kalmar K. Persistent postconcussion syn-
injury. J Int Neuropsychol Soc. 2005;11(3):228–36. drome: The structure of subjective complaints after mild
[PMID:15892899] traumatic brain injury. J Head Trauma Rehabil. 1995;10(3):
[Link] 1–17.
8. Belanger HG, Curtiss G, Demery JA, Lebowitz BK, Vander- 18. James W. The principles of psychology. New York (NY):
ploeg RD. Factors moderating neuropsychological outcomes H. Holt and Company; 1890.
following mild traumatic brain injury: A meta-analysis. J Int 19. Twamley EW, Jeste DV, Bellack AS. A review of cognitive
Neuropsychol Soc. 2005;11(3):215–27. [PMID:15892898] training in schizophrenia. Schizophr Bull. 2003;29(2):359–82.
[Link] [PMID:14552510]
9. Wehman P, Targett P, West M, Kregel J. Productive work [Link]
and employment for persons with traumatic brain injury: 20. McGurk SR, Twamley EW, Sitzer DI, McHugo GJ, Mueser
What have we learned after 20 years? J Head Trauma KT. A meta-analysis of cognitive remediation in schizo-
Rehabil. 2005;20(2):115–27. [PMID:15803036] phrenia. Am J Psychiatry. 2007;164(12):1791–1802.
[Link] [PMID:18056233]
10. Dawson DR, Schwartz ML, Winocur G, Stuss DT. Return [Link]
to productivity following traumatic brain injury: cognitive, 21. Wykes T, Huddy V, Cellard C, McGurk SR, Czobor P. A
psychological, physical, spiritual, and environmental corre- meta-analysis of cognitive remediation for schizophrenia:
lates. Disabil Rehabil. 2007;29(4):301–13. Methodology and effect sizes. Am J Psychiatry. 2011;
[PMID:17364780] 168(5):472–85. [PMID:21406461]
[Link] [Link]
68

JRRD, Volume 51, Number 1, 2014

22. Sitzer DI, Twamley EW, Jeste DV. Cognitive training in 32. Wehman P, Kregel J, Keyser-Marcus L, Sherron-Targett P,
Alzheimer’s disease: A meta-analysis of the literature. Acta Campbell L, West M, Cifu DX. Supported employment for
Psychiatr Scand. 2006;114(2):75–90. persons with traumatic brain injury: A preliminary investi-
[PMID:16836595] gation of long-term follow-up costs and program effi-
[Link] ciency. Arch Phys Med Rehabil. 2003;84(2):192–96.
23. Carney N, Chesnut RM, Maynard H, Mann NC, Patterson [PMID:12601649]
P, Helfand M. Effect of cognitive rehabilitation on out- [Link]
comes for persons with traumatic brain injury: A system- 33. Weathers FW, Keane TM, Davidson JR. Clinician-
atic review. J Head Trauma Rehabil. 1999;14(3):277–307. administered PTSD scale: A review of the first ten years of
[PMID:10381980] research. Depress Anxiety. 2001;13(3):132–56.
[Link] [PMID:11387733]
24. Koehler R, Wilhelm EE, Shoulson I; U.S. Institute of Medi- [Link]
cine Committee on Cognitive Rehabilitation Therapy for 34. Wilkinson GS. WRAT-3: Wide Range Achievement Test
Traumatic Brain Injury. Cognitive rehabilitation therapy for administration manual. 3rd ed. Wilmington (DE): Wide
traumatic brain injury: Evaluating the evidence. Washington Range Inc; 1993.
(DC): National Academies Press; 2011. 35. Raskin S. Memory for intentions screening test. J Int Neuro-
25. Cicerone KD, Langenbahn DM, Braden C, Malec JF, Kal- psychol Soc. 2004;10(Suppl 1):110.
mar K, Fraas M, Felicetti T, Laatsch L, Harley JP, Bergquist 36. Wechsler D. WAIS-III: Wechsler Adult Intelligence Scale.
T, Azulay J, Cantor J, Ashman T. Evidence-based cognitive 3rd ed. San Antonio (TX): Psychological Corp; 1997.
rehabilitation: Updated review of the literature from 2003 37. Delis DC, et al. CVLT-II: California Verbal Learning Test.
through 2008. Arch Phys Med Rehabil. 2011;92(4):519–30. 2nd ed. San Antonio (TX): Psychological Corp; 2000.
[PMID:21440699] 38. Delis DC, Kramer E, Kaplan JH. Delis–Kaplan Executive
[Link] Function System: Examiner’s manual. San Antonio (TX):
26. Comper P, Bisschop SM, Carnide N, Tricco A. A system- Psychological Corp; 2001.
atic review of treatments for mild traumatic brain injury. 39. Kongs SK. WCST-64: Wisconsin Card Sorting Test-64
Brain Inj. 2005;19(11):863–80. [PMID:16296570] card version; Professional manual. Odessa (FL): Psycho-
[Link] logical Assessment Resources; 2000.
27. Tiersky LA, Anselmi V, Johnston MV, Kurtyka J, Roosen E, 40. Blake DD, Weathers FW, Nagy LM, Kaloupek DG, Gusman
Schwartz T, Deluca J. A trial of neuropsychologic rehabilita- FD, Charney DS, Keane TM. The development of a Clinician-
tion in mild-spectrum traumatic brain injury. Arch Phys Med Administered PTSD Scale. J Trauma Stress. 1995;8(1):75–90.
Rehabil. 2005;86(8):1565–74. [PMID:16084809] [PMID:7712061]
[Link] [Link]
28. Management of Concussion/mTBI Working Group. VA/ 41. Hamilton M. Development of a rating scale for primary
DoD clinical practice guideline for management of concus- depressive illness. Br J Soc Clin Psychol. 1967;6(4):278–96.
sion/mild traumatic brain injury. J Rehabil Res Dev. 2009; [PMID:6080235]
46(6):CP1–68. [PMID:20108447] [Link]
29. Twamley EW, Vella L, Burton CZ, Heaton RK, Jeste DV. 42. Lehman AF. A quality of life interview for the chronically
Compensatory cognitive training for psychosis: Effects in a mentally ill. Eval Program Plann. 1988;11(1):51–62.
randomized controlled trial. J Clin Psychiatry. 2012;73(9): [Link]
1212–19. [PMID:22939029] 43. Bogdanova Y, Verfaellie M. Cognitive sequelae of blast-
[Link] induced traumatic brain injury: Recovery and rehabilitation.
30. Raskin SA, Buckheit CA, Waxman A. Effect of type of Neuropsychol Rev. 2012;22(1):4–20. [PMID:22350691]
cue, type of response, time delay and two different ongoing [Link]
tasks on prospective memory functioning after acquired 44. Vasterling JJ, Brailey K, Proctor SP, Kane R, Heeren T,
brain injury. Neuropsychol Rehabil. 2012;22(1):40–64. Franz M. Neuropsychological outcomes of mild traumatic
[PMID:22181940] brain injury, post-traumatic stress disorder and depression
[Link] in Iraq-deployed US Army soldiers. Br J Psychiatry. 2012;
31. Kreutzer JS, Wehman P, Morton MV, Stonnington HH. 201(3):186–92. [PMID:22743844]
Supported employment and compensatory strategies for [Link]
enhancing vocational outcome following traumatic brain
injury. Brain Inj. 1988;2(3):205–23. [PMID:3167277] Submitted for publication January 23, 2013. Accepted in
[Link] revised form July 11, 2013.
 69

TWAMLEY et al. CogSMART for Veterans with TBI

This article and any supplementary material should be Dev. 2014;51(1):59–70.

cited as follows: [Link]
Twamley EW, Jak AJ, Delis DC, Bondi MW, Lohr JB.
Cognitive Symptom Management and Rehabilitation
Therapy (CogSMART) for Veterans with traumatic brain
injury: Pilot randomized controlled trial. J Rehabil Res