THE PREDICTIVE AND INCREMENTAL VALIDITY OF THE FRONTAL
ASSESSMENT BATTERY (FAB) IN ON-ROAD DRIVING
PERFORMANCE AFTER BRAIN INJURY
David Dayton Hargrave
B.A., University of Colorado at Boulder, 2003
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
This thesis for the Master of Arts
David Dayton Hargrave
has been approved
Hargrave, David Dayton (M.A., Clinical Psychology)
The Predictive and Incremental Validity of the
Frontal Assessment Battery in On-Road
Driving Performance After Brain Injury
Thesis directed by Assistant Professor Elizabeth S. Allen
The question of fitness to drive often arises after a patient has suffered
a brain injury or been diagnosed with neurological disorder. Pre-driving
assessments are usually performed prior to on-road assessments but there is no
uniformity as to the instruments employed. One functional domain that has
been suggested to be critical to driving is executive functioning. The Frontal
Assessment Battery (FAB) is a brief screening instrument which seeks to
measure this function. The present study utilized archival data from records
of patients with traumatic brain injury, acquired brain injury, and neurological
disorders to examine the validity of the FAB to predict on-road driving
assessment outcome and to determine if it would uniquely predict on-road
driving assessment outcome when analyzed in conjunction with the Trail
Making Test Part B (TMTB). Simple logistic regression analysis showed that
for every unit decrease in performance on the FAB, a given patient was
significantly) more likely to fail the on-road driving test (OR = 1.60, p <
0.001. In addition, the FAB uniquely and significantly contributed to the
prediction of driving assessment outcome when analyzed in conjunction with
the TMTB. The present study provides evidence of a novel use of the FAB.
The predictive utility of the relatively brief FAB could be especially valuable
in helping to guide on-road driving evaluations as well as provide for
increasingly informed decision making for the patient considering the
generally high out-of-pocket expense of an on-road driving evaluation.
This abstract accurately represents the content of tl 11 T
recommend its publication.
Elizabeth S. Allen
This thesis is dedicated to my parents, who always encouraged me to reach for
the top and who instilled in me the confidence to set lofty goals and never
My thanks to all my committee members for their invaluable contributions
and enthusiastic support of this project. I would also like to thank Rey
Erickson for allowing the use of this data as well as the Medical Records staff
at Spalding Rehabilitation Hospital for their time spent redacting the data.
TABLE OF CONTENTS
1.1 The Frontal Assessment Battery..................................3
1.2 Previous studies using measures with similarities to the FAB....6
1.3 Trail Making Test Part-B........................................8
1.4 Purpose of the Study............................................10
2.1 Research Participants...........................................12
2.2.1 Pre-Driving Examination.......................................13
2.2.2 On-Road Driving Evaluation....................................14
2.3.1 The Frontal Assessment Battery................................15
2.3.2 The Trail Making Test Part-B..................................16
2.4 Statistical Analysis...........................................17
4.3 Future Studies...............................................28
B. Human Subject Approval........................................36
LIST OF TABLES
3.1 Logistic regression predicting on-road driving performance from FAB
3.2 Logistic regression predicting on-road driving performance from FAB
3.3 Logistic regression predicting on-road driving performance from FAB
and TMTB score..................................................23
Traumatic brain injury (TBI), acquired brain injury (i.e.
cerebrovascular accident), and various neurological disorders often result in
cognitive impairments which interfere with safe driving behaviors (McKenna,
Jefferies, Dobson, and Frude, 2004; Yale, Hansotia, Knapp, and Ehrfurth,
2003). Research suggests that 60% of brain injury survivors return to driving
and drive at least 50 miles every week; as many as 66% of them do so without
formal evaluation (Brower & Withaar, 1997; Fisk, Schneider, and Novack,
1998). Within a rehabilitation setting, fitness to drive following brain injury is
consistently one of the most salient issues encountered (Tamietto et al., 2006).
Because driving is critical to independence and community reintegration,
patients often report the possibility of driving restrictions to be their greatest
functional concern subsequent to injury (Rapport, Hanks, and Bryer, 2006).
Evaluations to assess fitness to drive usually include pre-driving screening
instruments and an on-road driving assessment; however, there is currently no
accepted standard in evaluating fitness to drive after brain injury (Komer-
Bitensky, Bitensky, Sofer, Man-Son-Hing, and Isabelle, 2006; Tamietto et al.,
2006;). The current study seeks to explore the value of the predictive ability
of the Frontal Assessment Battery (FAB) alone and in combination with the
Trail Making Test Part B (TMTB) in on-road driving performance in a sample
of patients with brain injury.
A component of driving assessments which reflects the high degree of
variability among evaluation methods is in the selection of pre-driving
screening instruments. Yale et al. (2006) suggested that there are at least 44
commonly used neuropsychological instruments to assess cognitive
impairments related to driving fitness. Indeed, the research is highly variable
with regard to the instruments employed and which outcome measures are
used as validators (Schultheis, Matheis, Nead, and Deluca, 2002). For
example, some studies have used driving simulators whereas others have used
fixed-course off-road tests to validate the predictive utility of
neuropsychological tests (Korteling & Kaptein, 1996; Lundqvist et al., 1997;
Lundqvist, Gerdle, and Ronnberg, 2000). Collectively, results of studies
investigating the value of neuropsychological test scores in predicting
subsequent on-road driving performance have been mixed and no instrument
or group of instruments has been consistently replicated as significantly
predictive of driving fitness after brain injury (Tamietto et al., 2006).
It has been suggested that one problem with the extant literature on
this topic is the relative lack of a thorough assessment of executive function as
a predictor of driving ability (Coleman et al., 2002; Tamietto et al., 2006). By
and large, studies examining the usefulness of pre-driving screening
instruments in brain injury populations have focused on basic visual,
perceptual, and motor functioning (Akinwuntan et al., 2006; Marshall et al.,
2007; Soderstrom, Petterson, and Leppert, 2006; Tamietto et al., 2006).
Executive function refers to higher-order, goal-directed cognitive processes
including those related to planning, problem solving, initiation, inhibition,
cognitive flexibility, and deficit awareness. Although intact gross
visuoperceptive ability (e.g., absence of neglect or severe inattention), intact
short and long-term memory, and sufficient motor coordination and speed are
required for safe driving, intact executive functioning may be the most salient
issue in determining fitness to drive after brain injury because it allows for the
integration of numerous processes critical to safe driving. Thorough
assessment of this type of functioning may play an important role in tailoring
on-road assessments for a given individual, the safety of the evaluator in such
assessments, and formulation of ongoing rehabilitation strategies.
1.1 The Frontal Assessment Battery
The Frontal Assessment Battery (FAB) is a recently developed brief
instrument intended to assess executive functioning (Dubois, Slachevsky,
Litvan, & Pillon, 2002). The FAB consists of six items designed to measure
abstraction, lexical fluency, motor programming, response interference,
inhibition, and prehension behavior. Abstraction is a measure of cognitive
flexibility, or the ability to adapt cognitive processing to accommodate novel
conditions in ones environment. This function is measured via similarities
(e.g., How are a banana and an orange alike?) on the FAB; overly concrete
responses are thought to reflect a deficit in problem solving approach. Lexical
fluency is a general measure of ones ability to organize thought and
formulate strategies towards reaching a goal. The lexical fluency task requires
that the patient produce as many words as possible beginning with the same
letter in one minute. Motor programming is measured by asking the patient to
perform a sequence of three motor movements (fist-edge-palm). This gives
the evaluator a measure of the patients ability to link sequential motor
movements in a fluid and meaningful fashion. Response interference is
measured by asking the patient to perform a finger tapping exercise in concert
with the evaluator. In this task, the patient is asked to tap once when the
evaluator taps twice and to tap twice when the evaluator taps once. An
inability to perform this task is indicative of an inadequate ability to block out
interfering stimuli. Response inhibition is measured by asking the patient to
tap once every time the examiner does so and to do nothing when the
examiner taps twice. The patients score on this item is indicative of their
ability to inhibit incorrect behavioral impulses. Prehension behavior is
measured by the examiner placing their hands above the patients palms to see
if the patient will spontaneously take the examiners hands. If the patient does
so, they are given another trial and told do not take my hands. This is a
measure of ones ability to function autonomously within their environment
and not of stimulus bound nature.
The functions tapped by the FAB could be argued to be reflective of
abilities and functions necessary for safe driving. For example, overly
concrete thought and highly stimulus-bound behavior could result in an
individual stopping in an intersection when the light turns red, rather than
completing their intended left-hand turn. Likewise, the inability to block-out
interfering stimuli (e.g. a conversation with a passenger when obstacles are
present in the road) or inhibit certain responses (e.g. sudden braking or
steering on an icy road) could have detrimental consequences while driving.
Finally, the inability to adapt to changes in a driving route (e.g. unexpected
construction or unanticipated presence of one-way streets) could result in
some individuals becoming lost or confused, potentially with dangerous
In variety of patient populations, the FAB correlates well with other
measures of executive function and general cognition (Dubois, et al., 2000;
Lipton et al., 2005; Matsui et al., 2006; Paviour et al., 2005; Slachevsky et al.,
2004). In addition, the FAB has been shown to be able to discriminate
between patients with frontal deficits and healthy control subjects and as a
result may be useful for patients with TBI, acquired brain injury (i.e.,
cerebrovascular accident), and various neurological disorders (Dubois, et al.,
2000; Royall, Lauterbach, Cummings, Reeve, and Rummans, 2002).
Currently, it is commonly used by rehabilitation personnel including
neurologists, neuropsychologists, and neuropsychiatrists. However, it has not
yet been used to predict driving ability and, given the central role of executive
function in the act of driving, it may prove to be a particularly useful measure
in the assessment of this function.
1.2 Previous studies using measures with similarities to the FAB
Although the FAB has never been used for the purpose of predicting
driving outcome, studies utilizing measures of executive functioning which
tap sub-domains similar to those measured by the FAB have produced
promising results For example, the Wisconsin Card Sorting Task (WCST),
which measures cognitive flexibility/problem-solving, has previously been
used in this fashion (Lezak, 2003). Lundqvist et al. (1997) found a significant
difference in the percentage of WCST perseverative errors between stroke
patients and controls in terms of risk awareness in on-road driving
performance with low risk awareness being correlated with failure on the on-
road driving test.
Lundqvist et al. (1997) found that performance on the Complex
Reaction Time test, a measure of response inhibition, was the most predictive
of on-road driving failure, correctly classifying 83% of the subjects. This last
finding is important because almost no studies of neuropsychological driving
prediction include a pure measure of response inhibition. However, the FAB
does measure this sub-domain and the finding of Lundqvist et al. (1997) offers
support for its use.
The Stroop task, in its numerous variants, is a well-established
measure of response interference, another sub-domain tapped by the FAB
(Lezak, 2003). Its use in the driving literature appears to be uncommon even
among studies which examine executive function. In a 2004 study, Radford,
Lincoln, and Murray-Leslie found that poor performance on the Stroop
Neuropsychological Screening Test by TBI patients was strongly correlated
with failure on a subsequent on-road driving evaluation. In an interesting
study examining how normal individuals handle a critical driving situation,
Collet, Petit, Priez, and Dittmar (2005) found that good performance on the
Stroop task was predictive of subjects who would perform adequate
maneuvering when an unexpected obstacle was thrown into their path while
driving. These studies offer support for the response interference portion of
In 2004, Mckenna et al. carried out a driving study with a mixed group
of neurologically compromised patients in which they investigated the
predictive validity of two subtests which are identical to two of the six FAB
subtests. The subtests, which were derived directly from Lurias (1973) (as
cited in Mckenna et al. 2004) work and are analogous to the response
interference and fist-edge-palm subtests of the FAB, were included in a
test battery which correctly classified 100% of participants aged 69 years and
younger and 84% of participants aged 70 years and older as those who would
fail the on-road driving evaluation. Only 5.5% of control patients failed the
cognitive test battery. These findings add further support for the validity of
the FAB in the prediction of on-road driving.
1.3 The Trail Making Test Part-B
Another neuropsychological test which measures specific executive
functions and has been shown to be useful in driving assessment is the Trail
Making Test Part B (TMTB). The TMTB task is designed to measure set
shifting (i.e., alternating between different rule sets necessary for the
completion of a task), alternating attention (i.e., switching attention between
two different tasks), visual scanning (e.g., ability to find salient items in ones
environment), and sequencing (i.e., planning and programming). It requires
that the subject alternate between sequential numbers and letters by drawing
lines between them (see methods section for more details).
In a recent comprehensive review of studies utilizing samples of stroke
survivors, Marshall et al. (2007) found that the TMTB was one of the most
predictive tests of cognitive functioning in terms of on-road driving
performance after stroke. Soderstrom, Petterson, and Leppert (2006) found
that stroke patients performed significantly worse on the TMTB than did
controls. In addition, in a sample of TBI patients, a study using the TMTB
found that this instrument was significantly predictive of poor risk awareness
in the on-road driving test and that these individuals were more likely to fail
the on-road evaluation (Lunqvist et al., 1997). Another study with TBI
patients found that performance on the TMTB positively correlated with
failure of a subsequent on-road driving test (Schanke & Sundet, 2000). A
study using the Colored Trails Test (CTT) Part B, which purports to measure
the same sub-domains as the TMTB, found performance on the CTT to be
predictive of post-TBI driving safety, with those performing poorly having a
significantly higher likelihood of automobile accident after injury (Coleman et
al., 2002). Finally, in support for use of the TMTB in patients with
neurological disorders, Stolwyk, Charlton, Triggs, Iansek, and Bradshaw
(2006) recently demonstrated that poor performance on the TMTB was
significantly correlated with on-road driving impairment in patients with
In combination, the FAB and TMTB may represent useful screening
instruments with regard to executive function. Although the TMTB has been
used in the past to determine fitness to drive in brain-injured populations, the
FAB has never been utilized in this fashion (Yale et al., 2003). Due to the
breadth of the brief FAB, its use in driving evaluations in this population may
prove significantly predictive in determining which patients may fail an on-
road assessment. When employed in combination with the TMTB test in an
identical context, the FAB may add to the value of these tests in predicting on-
road assessment outcome.
1.4 Purpose of the Study
The current study seeks to explore the value of the predictive ability of
the FAB alone and in combination with the TMTB for on-road driving
performance in a brain-injured population. This study will utilize archival data
from a hospital-based driving assessment and rehabilitation program to
investigate two primary questions. First, the independent relationships
between both FAB and TMTB scores and subsequent on-road driving
performance will be examined. It is hypothesized that poorer scores on the
FAB and TMTB will each predict a greater likelihood of receiving a failing
score in on-road driving assessment. Secondly, FAB scores will be combined
with scores from the TMTB to predict subsequent on-road driving
performance. It is hypothesized that scores from the FAB will add to the
predictive value of the TMTB and vice versa. Specifically, it is expected that
the FAB and TMTB will each demonstrate unique predictive ability in terms
of on-road driving performance.
2.1 Research Participants
Out of approximately 200 patients, 99 patients between 18 and 88
years of age (M 58.7, SD = 16.8, Mdn = 59), with whom driving
assessments were performed as a part of the Driving Rehabilitation Program
at Spalding Rehabilitation Hospital (SRH; Aurora, CO), were included in the
present study. A majority of the patients (66%) fell between the ages of 32
and 69. Participants were selected based on the presence of TBI (n = 29),
acquired brain injury (n = 47), or neurological disorder (anoxic brain injury,
Multiple Sclerosis, Amyotrophic Lateral Sclerosis, tumor, Parkinsons
disease, Alzheimers dementia, and vascular and undifferentiated dementia; [n
= 23]) as well as presence of complete assessment scores. The remaining
patients were excluded due to lack of a neurological diagnosis (i.e., patients
who were referred to the driving program as a result of recent amputations,
spine surgery, or mild cognitive impairment.) or incomplete data (i.e.,
lacking data from pre-screening or on-road evaluation. In addition, patients
who failed the pre-driving screening for intact basic visual, perceptual, and
motor abilities were never tested using the FAB or TMTB and, as a result,
were excluded from the present study. The sample was 74.7% (n = 74) male
and 25.3% (n = 23) female. Of the patients for whom education data was
available (n = 92), the mean education in years was greater than a high school
education (M = 14.68).
The Driving Rehabilitation Program at SRH evaluates outpatients
whose physicians are concerned about the patients ability to drive. The
program assesses these patients competence to drive and indicates
rehabilitation if necessary. Only data collected during the initial pre-driving
and on-road evaluations, before any rehabilitation took place, were included in
the present study. Participants whose medical record indicated recent
comorbid diagnoses (e.g. TBI and CVA) were excluded. However,
participants whose medical record indicated previous diagnoses (eg. multiple
TBIs) were included.
2.2.1 Pre-Driving Examination
The first part of the pre-driving examination consisted of an
assessment of visual acuity, contrast sensitivity, color and depth perception,
lateral and vertical phoria (e.g., the position of the eyes when binocularly
fixed on a given object), peripheral vision, reaction time, road sign
recognition, and visual field perception. The Certified Driver Rehabilitation
Specialist (CDRS; Registered Occupational Therapist) examined visual
acuity, contrast sensitivity, color and depth perception, lateral and vertical
phoria and peripheral vision using the OPTEC 2000 machine. Visual field
perception was tested via the Acuvision 1000, reaction time via the Reaction
Time Visual Scanning Test, and road sign recognition via the Road Sign
Recognition Test. After this initial assessment, the examination was
continued by the CDRS under the supervision of a Licensed Clinical
Psychologist. If the patient successfully passed the above visual, perceptual,
motor screening, the CDRS then administered the FAB and the TMTB
measures to the patient. The CDRS then scored the measures and forwarded a
report to the supervising psychologist for review and interpretation.
The CDRS received training by the supervising psychologist in
administration and scoring procedures of both instruments. Administration
and scoring techniques for the FAB were derived from Dubois et al. (2000)
and administration and scoring procedures for the TMTB were derived from
2.2.2 On-Road Evaluation
After the pre-driving examination, the Adaptive Driving Evaluation
(ADE) was completed, usually within three weeks. The ADE consists of a
standard driving route that generally lasts for approximately 28 miles and
includes driving in residential traffic, city traffic, high-volume traffic, and the
expressway. The same automatic transmission test car was employed in all
evaluations. If significant safety concerns arise during the course of any given
evaluation, the test may be attenuated and not include all portions of the
standard route. The ADE includes assessment of all technical skills pertinent
to driving including use of all vehicle instrumentation and adjustments
necessary prior to driving. The stated goals of the ADE are that the patient be
able to consistently perform the following functions, with no physical or
verbal assistance from the evaluator, in all traffic conditions: (1) safe entry
and exit from intersections, (2) demonstration of sufficient attention and
concentration, (3) acceptable visual scanning behaviors, (4) execution of safe
lane changes including over-the-shoulder checks, and (5) complete stops at all
stop signs. The ADE concludes with an informal assessment of the patients
insight into their driving ability. The evaluator determines a pass or fail
outcome based primarily on the five aspects of safe driving listed above.
2.3.1 The Frontal Assessment Battery
The FAB is a six item test designed to measure various aspects of
executive functioning (abstraction, response inhibition, interference, motor-
programming, prehension, and mental flexibility) and to be completed within
10 minutes. The FAB has been shown to significantly discriminate between
patients with frontal lobe dysfunction and healthy controls by correctly
identifying 89.1% of cases in a sample of 121 patients and 42 controls
(Dubois et al., 2000). Furthermore, the similarities and prehension subtests of
the FAB were able to discriminate between patients with cortical versus
subcortical dysfunction, correctly identifying 69.7% of cases (Dubois et al.,
2000). The FAB has shown concurrent validity in a sample of patients with
various frontal disorders when compared with scores on the Mattis Dementia
Rating Scale (MDRS) (r = 0.82, p < .001) on both criteria achieved (r = 0.77,
p < 0.001) and number of perseverative errors on the Wisconsin Card Sorting
Task (WCST) (rho = 0.68,/? < 0.001) (Dubois et al., 2000). Interrater
reliability was significant when evaluated with 17 patients from the same
population (k = 0.87,/? < 0.001) and internal consistency was adequate with
scores from all 121 patients (r = 0.78) (Dubois et al., 2000).
2.3.2 Trail Making Test Part B
The TMTB is a well-established brief instrument designed to measure
divided attention, planning ability, and set shifting (the ability to switch
between rule sets of different tasks). The test requires that the subject connect
letters and numbers in a sequential and alternating fashion by drawing lines
between them; the subject is directed to complete this task as fast as you
can. The measure, which begins with a sample test orienting the participant
to the task procedure, has on the front side a box containing numbers and
letters which are each enclosed by circles. The participant is asked to draw a
line from the number 1 to the letter A and from A to 2, from 2 to
B and so on, always alternating between a number and a letter in sequential
order. After correctly completing the sample portion, the participant is asked
to complete one formal trial which consists of a longer version of the sample
test, containing numbers from 1-13 and letters from A-L. The TMTB is
scored by measuring the time necessary for the participant to complete the
In her 2003 review, Lezak states that performance on the TMTB is
highly vulnerable to brain damage of all types including mild head trauma
and the results of some studies lend support to a role for frontal brain regions
in the task. In addition, the task has been shown to be sensitive to the
cognitive effects of various types of dementia (Lezak, 2003). Test-retest
reliability ranged from 0.60 to 0.90 depending on the study (Lezak, 2003).
However, for a task such as the TMTB, the issue of practice effects may be an
2.4 Statistical Analysis
Data was analyzed using SPSS for Windows (ver. 16). Usually, higher
scores on the FAB are indicative of better performance, but for these analyses
scores from the FAB were inverted to yield a more intuitive odds ratio
statement. Also, to yield a range of scores more similar to that of the FAB,
TMTB raw scores were converted to scaled scores based on Heatons (2004)
normative data. Like the FAB, a higher scaled score on the TMTB is
indicative of better performance; these scaled scores were inverted as well to
yield a more intuitive odds ratio statement. The range of possible scores on
the FAB was 0-18 and the range of possible TMTB scaled scores was 1-19.
All hypotheses were tested with binomial regression to evaluate how
the scores on the FAB and TMTB predicted the odds of passing or failing on
the driving test. Specifically, scores from the FAB were entered into a simple
binomial logistic regression equation to test the hypothesis that scores on the
FAB can predict on-road driving assessment outcome (e.g., pass/fail). Next,
the value of TMTB scores to predict on-road driving performance was
evaluated with simple binomial logistic regression. Finally, scores from the
TMTB and FAB were entered simultaneously in a multiple binomial logistic
regression to test the hypothesis that the FAB and TMTB would each
demonstrate unique predictive ability.
Odds ratios that diverge significantly from a value of 1.0 in either
direction indicate significant increases or decreases in the odds of a given
outcome (e.g., passing the driving test). Significance was set at p < .05 and
tested with the Wald statistic. This post-hoc test of significance seeks to
determine whether the experimental equation model differs significantly from
a null model.
Participants raw scores (before inversion) on the FAB ranged from 7-
18 and TMTB scaled scores (before inversion) ranged from 1-15. Participants
who obtained a raw score on the FAB worse than seven were not tested on
road due to the severity of their impairment. This determination was made on
an individual basis by the CDRS performing the evaluation.
Both simple and multiple logistic regression analyses were performed
to determine the extent to which the independent variables predicted on-road
driving evaluation outcome independently and when entered simultaneously.
In a simple logistic regression equation the FAB was shown to be a significant
predictor of on-road driving evaluation outcome, indicating that for every
point decrease in performance on the FAB an individual was 1.60 times more
likely to fail the on-road driving evaluation ip < .001; see Table 3.1). In this
same equation, total weighted correct classification of patients who would
either fail or pass the on-road driving assessment rose from 57.6% in the null
model to 73.7% in the model using the FAB. Frequency analysis
demonstrated that participants who obtained a raw score of 15 or worse (n =
47) on the FAB failed the on-road driving evaluation 83% of the time. In
contrast, patients who obtained a FAB raw score of 16 or better (n = 52) failed
the on-road driving evaluation only 35% of the time.
Table 3.1 Logistic regression predicting on-road driving performance from
95% Cl for exp b
B (SE) Lower exp b Upper
FAB .471*** 1.28 1.60 2.00 (.114)
Note Rl .22 (Cox & Snell), .29 (Nagelkerke). Model %2 (1) = 24.42, p< .001.
The TMTB was also shown to be a significant predictor of on-road
driving evaluation outcome, indicating that for every point decrease in
performance on the TMTB (using scaled scores) an individual was 1.52 times
more likely to fail the on-road portion of the driving exam (p < .001; see Table
3.2). In this equation, total weighted correct classification of patients who
would either fail or pass the on-road driving assessment rose from 57.6% in
the null model to 71.7% in the model using the TMTB.
Table 3.2 Logistic regression predicting on-road driving performance from
95% Cl for exp b
B (SE) Lower exp b Upper
TMTB .420*** 1.25 1.52 1.85 (.099)
Note Rz = .23 (Cox & Snell), .31 (Nagelkerke). Model %2 (1) = 25.79,p< .001.
When scores from both instruments were entered simultaneously into a
multiple logistic regression equation both FAB and TMTB scores made
significant, unique contributions in predicting failure of the on-road driving
evaluation. In this model the FAB yielded an adjusted odds ratio of 1.36 (p <
.05) while the TMTB yielded an adjusted odds ratio of 1.34 (p < .01) (see
Table 3.3). Thus, when the FAB and TMTB were analyzed together, for
every point decrease performance on the FAB a given participant was 1.36
times more likely to fail the on-road driving evaluation while for every point
decrease in performance on the TMTB (using scaled scores) a given
participant was 1.34 times more likely to fail the on-road driving evaluation.
In this model, total weighted correct classification of patients who would
either fail or pass the on-road driving assessment rose from 57.6% in the null
model to 72.7% in the model using both the FAB and TMTB.
Table 3.3 Logistic regression predicting on-road driving performance from
TAB and TMTB score.
95% Cl for exp b
B( SE) Lower exp b Upper
FAB .306* 1.06 1.36 1.744
TMTB .292** 1.08 1.34 1.67
Notei?2= .28 (Cox & Snell), .37 (Nagelkerke). Model ^(2) = 32.15,p< .001.
* p<.05. ** p <.01.
The present study sought to determine the predictive validity of the
FAB and TMTB independently and combined in on-road driving evaluation
outcome subsequent to diagnosis of brain injury or neurological disorder. The
FAB was selected as a measure of executive functioning due to its breadth,
brief format, and its novel use in this population while the TMTB was selected
as an additional measure of executive functioning due to its generally
established credentials and ability to tap domains of executive functioning that
the FAB cannot. Both the primary and secondary hypotheses were supported
by the findings. Logistic regression analyses demonstrated the FAB and
TMTB to both have significant predictive power independently and, when
combined, to each retain a significant unique contribution to predictive power
in on-road driving assessment outcome.
The findings of the present study could result in a more informed
decision making process for both patients and clinicians when considering on-
road driving assessment. Comprehensive driving assessments which include
an on-road component are usually paid out-of-pocket and are relatively
expensive. Thus, if individual patients could be provided with data
demonstrating their likelihood of on-road failure, they may elect to defer this
component of the evaluation in lieu of further rehabilitation or may decide to
forgo this portion all together. For example, the information that out of
approximately 50 patients, those who obtained a raw score of 15 or worse on
the FAB failed the on-road evaluation 83% of the time could substantially
change a low-scoring patients decision options. Thus, informed decision
making could not only save the patient time and money but would save the
evaluators time for patients who are more likely to pass this exam. In
addition, deficits elucidated by the FAB could help guide the on-road
evaluation to maximize both the evaluators safety and further delineate how a
given patients pre-driving test performance may correlate with real-world
driving ability. For instance, using data gleaned from an FAB profile, the
evaluator could choose to place the patient in progressively more taxing
situations with regard to executive abilities while simultaneously being able to
anticipate a patients response to a given stimulus. Not only could this result
in a more comprehensive and useful evaluation but it may help to ensure the
safety of both the evaluator and the patient.
The results of the present study should not be construed such that, in
the assessment of executive function in pre-driving screening, clinicians rely
solely on the FAB and TMTB. Rather, the present findings suggest that the
FAB may be a useful measure to complement batteries of executive function
measures currently being used by driving assessment programs. The brevity
of the administration of this instrument bolsters this idea for adding ten
minutes to the completion time of an assessment battery is unlikely to be
problematic for clinicians or patients. However, replication of the results of
the present study is needed to further clarify the validity of the FAB in the
prediction of on-road driving performance after brain injury.
The current study utilized a diagnostically mixed group of patients,
consistent with most earlier literature and with common clinical practice (e.g.,
applying a standard battery to diverse patients). However, some predictive
driving studies have separated groups of participants by diagnostic category or
lesion location (Bouillon, Mazer, and Gelinas, 2006; Tamietto et al., 2006).
Future research could extend this study by testing the predictive ability of the
FAB for driving in specific subpopulations. However, our findings that an
instrument such as the FAB demonstrates predictive validity across a diverse
group of neurologically compromised patients is generally more amenable to
the way driving assessments are actually conducted. For example, it would
likely be unreasonable in logistical terms for a driving rehabilitation program
to develop a separate battery of tests for each patient diagnostic category. The
cost of stocking these tests and training staff members on the nuances of the
administration of each would be unreasonable and likely raise the cost of this
already expensive process for the patient.
The archival nature of this study brings forth limitations that warrant
further discussion. First, the majority of the patients included in this study
were referred to the driving rehabilitation program from outside providers. As
a result, this may have biased the sample away from those individuals with
more serious injuries. Although complete data of this nature was not
available, it can be assumed that because most of the patients in this study
originated from outside referrals they were relatively independent in terms of
functional status at the time of evaluation.
Second, the relatively high education level of the participant sample
may have resulted in overall better performance on the test measures.
Appollonio et al. (2005) demonstrated that FAB scores can vary as a function
of education. Thus, the education level of the sample used in the current study
(14.68 years on average) may have resulted in inflated scores across the
sample. In a sample with an average of 12 years of education the outcome
may be different.
Third, since driving evaluation is relatively expensive and is generally
paid out of pocket, the sample used in the present study may have been biased
towards patients with higher socioeconomic status (SES). Researchers often
produce normative data that is corrected for this factor as a result of the fact
that test performance can vary as a function of SES. Data of this nature was
not available in the current study.
Fourth, data on transient conditions such as the effect of current
medications or pain levels could have allowed the ability to screen out certain
participants or altered the testing procedure such that the participants would
be tested in the absence of these conditions. It is known that the experience of
significant pain or the side-effects of certain drugs can temporarily alter an
individuals cognition. Thus, information of this nature could be vital to
producing a valid testing profile both during pre-driving clinical assessment
and during on-road evaluation. This data was not available in the present
4.3 Future Studies
Future studies should seek to replicate the results of the present study
demonstrating the predictive validity of the FAB screening instrument in on-
road driving performance after brain injury. In addition, future studies would
benefit from the ability to analyze the potential interaction effects of other
types of data including ethnicity, gender, and SES, among others.
The possibility of gender differences in performance on measures of
executive functioning should be addressed in future studies. Other researchers
have suggested that there may be an effect of gender on tests of executive
functioning after brain injury (Niemeier, Marwitz, Lesher, and Walker, 2007).
Thus, it might be that females would perform better overall than males on the
Furthermore, a record of actual driving infractions or collision data, in
the form of Department of Motor Vehicles records, could be valuable in future
studies in two ways. First, if collected before evaluation, it could provide an
indication of premorbid driving function. Second, this same data collected
after brain injury and before or after testing could provide an additional
functional correlate of the FABs predictive power.
Future investigation should also examine the role of insight in the
moderation of executive functioning deficits. It may be that an individuals
level of deficit awareness could act in a protective fashion. For example, an
individual with significant deficits in executive functioning and awareness of
these deficits may choose to drive within a predefined restricted area that
would not extensively tax their cognitive reserves, resulting in safer driving
for the impaired driver.
In conclusion, both the primary and secondary hypotheses of the
present study were supported, demonstrating that the FAB is a significant
predictor of on-road driving performance after brain injury and that this
property is retained when scores from the FAB and TMTB were evaluated
within the same equation. These findings add to the growing body of
literature indicating the value of measures of executive for predicting driving
ability after brain injury. The findings offer support for the use of a screening
measure that may be valuable for the patient and the clinician in guiding on-
road driving assessment and dictating on-going rehabilitation strategies.
Future investigation should seek to replicate the present study and expand on
it by evaluating additional measures of executive functioning in concert with
the FAB, using different patient populations, and investigating potential
interaction effects of moderating variables. The results of the present study
and previous studies evaluating similar domains of executive functioning
suggest that future studies seek to elucidate the optimal combination of tests
of executive functioning to provide a battery that is most predictive of on-road
driving ability after brain injury. It is likely that the optimal combination of
measures of executive functioning for this purpose would include instruments
such as the FAB, TMTB, WCST, the Stroop task, and the Complex Reaction
(5) C ftv
Frontal Assessment Battery
I. Similarities (conceptualization)
' In what ways arc they alike?"
A banana and an orange,.
(In the event of total failure: They are not alike" or partial failure. 'Both have a peel" help the patient by saving:
Both a banana ami an orange are...", but credit 0 for the item, ana do not help the patient for the two following
A rabie and a chair
A tulip, a rose, and a daisy
__ Score; (only category responses fruits, furniture, flowers arc considered correct)
Three correct: 3
Two correct; 2
One correct: 1
None correct: 0
1. l exical Fluency (mental flexibility)
Say as many words as you can beginning with the letter S... any words except surnames or proper nouns,"
If the patient gives no response during the first five seconds, say: ''for instance, snake. If the patient pauses \0
seconds, stimulate him by saving: any word beginning with the letter S. The time allowed is *0 seconds
Score: (word repetitions or variations [shoe, shoemaker], surnames, or proper nouns are not
counted as correct responses.
More than nine words: 1
Six *r> nine word* 3
Three to five words: l
Less than three words: 0
3. Motor Series (programming)
"Took carefully at what I'm doing.
The examiner, seated in front of the patient, per forms three times alone with his left hand the senes of l unu fis
edge-palm." Now. with your right hand, do the same series, first with me and then alone. The examiner pci form
the scries three times with the patient, then says to him her: New do it on your own.
Patient performs six correct consecutive series alone: }
Patient performs at least three correct consecutive series alone: 2
Patient fails alone, but performs three correct consecutive series with the examiner: I
Patient cannot perform three consecutixe even with the examiner; 0
4. ( onflicting Instructions (sensitivity to interference)
Tap twice when 1 tup once.
Be sure that the patient has understood ihe instruction, a senes of three trials is nan: I -1-1.
' Tap once when i tap twice."
To be sure that the patiem lias understood the instruction, a senes of three trials is run 2-2-2.
The examiner performs the following senes 1-1-2 -1-2-2-2 -1-1-2
No error- 3
One or two errors: 2
More than two erroi s I
Patient taps like the examiner at four consecutive mnes 0
Krontal Assessment Battery (tout.)
5. Go No Go (inhibitory control)
Tap once when I tap once.
\ o he sure that the patient has understood the instruction, a senes of three trials is tun; N ]-|.
T)o not tap when up twice.** To be sure that the pattern has understood the instruction, a series of throe trials i<
The examiner performs the following senes. I -1-2-1 -2-2-2-1 i -2
No error: 3
One or two tutors: 2
Y!ore than two errors: I
Patient taps like the examiner at least four consecutive times. 0
6. Prehension Behavior (environmental autonomy)
Do not take my hands.*
The examiner is seated in front of the patient. Place the patient 's hands palm up on his/her knees Without sayim
anything or looking at the patient, the examiner brings his-her hands close to the patients hands and touches the
palms of both the patient's, to see if he/she will spontaneously take them. If the patient takes the hands, the
examiner w ill try again after asking him her; Now. do not take mv hands."
Patient does not tale ihe examiner s hands: 3
Patient hesitates and asks what he. she has to do: 2
Patient takes the hands without hesitation: I
Patient takes the examiners hands even after he she has been told not to do so: 0
/18 Total Score
HUMAN SUBJECTS APPROVAL
University oe Colorado
AT 'MW- --N'T Hr* T'- A "A r: MII .
Certificate of Exemption
Investigator: David Hargrave
Subject HSRC Protocol 2008-087
Initial Review (APP001;
14 November 2007
THE PREDICTIVE AND INCREMENTAL VALIDITY OF THE FRONTAL
ASSESSMENT BATTERY (FAB) IN ON-ROAD DRIVING PERFORMANCE
AFTER BRAIN INJURY
This protocol qualifies for exempt status. Periodic continuing review is not required. For the duration of your
protocol, any change in the experimental design/content of this study must be approved by the HSRC before
implementation of the changes.
The anticipated completion date of this protocol is 11Â£14Â£2010. HRSC will terminate this project on this date unless
otherwise instructed either by correspondence, telephone or e-mail
'//( / .............._
Akinwuntan, A.E., Feys, H., De Weerdt, W., Baten, G., Amo, P., and
Kiekens, C. (2006). Prediction of Driving after Stroke: A Prospective
Study. Neurorehabilitation and Neural Repair, 20, 417-423.
Alvarez, J.A. & Emory E. (2006). Executive Function and the Frontal Lobes:
A Meta Analytic Review. Neuropsychology Review, 16(1), 17-42.
Appollonio, I., Leone, M., Isella, V., Piamarta, F., Consoli, T., Villa, M.L.,
Forapani, E., Russo, A., and Nichelli, P. (2005). The Frontal
Assessment Battery (FAB): normative values in an Italian population
sample. Neurological Science, 26, 108-116.
Bouillon, L., Mazer, B., and Gelinas, I. (2006). Validity of the Cognitive
Behavioral Drivers Inventory in Predicting Driving Outcome. The
American Journal of Occupational Therapy, 60(4), 420-427.
Brower, W.H. & Withaar, F.K. (1997). Fitness to Drive After Traumatic Brain
Injury. Neuropsychological Rehabilitation, 7(3), 177-193.
Coleman, R.D., Rapport, L.J, Ergh, T.C., Hanks, R.A., Ricker, J.H., and
Millis, S.R. (2002). Predictors of Driving Outcome After Traumatic
Brain Injury. Archives of Physical Medicine and Rehabilitation, 83,
Dubois B., Slachevsky A., Litvan I., and Pillon B. (2000). The FAB: a
Frontal Assessment Battery at bedside. Neurology, 55, 1621-1626.
Fisk, G.D., Schneider, J.J., and Novack, T.A. (1998). Driving following
traumatic brain injury: prevalence, exposure, advice, and evaluations.
Brain Injury, 12(8), 683-695.
Heaton, R.K., Miller, S.W., Taylor, M.J., and Grant, I. (2004). Revised
comprehensive norms for an expanded Halstead-Reitan battery.
Odessa, FL: Psychological Assessment Resources.
Komer-Bitensky, N., Bitensky, J., Sofer, S., Man-Son-Hing, M., and Isabelle,
G. (2006). Driving Evaluation Practices of Clinicians Working in the
United States and Canada. The American Journal of Occupational
Therapy, 60(4), 428-434.
Lezak, M.D. (2003). Neuropsychological Assessment (3rd. ed.). New York:
Lipton, A.M., Ohman, K.A., Womack, K.B., Hynan, L.S., Ninman, E.T., and
Lacritz, L.H. (2005). Subscores of the FAB differentiate
frontotemporal lobar degeneration from AD. Neurology, 65, 726-731.
Lundqvist, A., Alinder, J., Aim, H., Gerdle, B., Levander, S., and Ronnberg, J.
(1997). Neuropsychological Aspects of Driving After Brain Lesion:
Simulator Study and On-Road Driving. Applied Neuropsychology, 4(4),
Lunqvist, A., Gerdle, B., and Ronnberg, J. (2000). Neuropsychological
Aspects of Driving After a Stroke- in the Simulator and on the Road.
Applied Cognitive Psychology, 14, 135-150.
Marshall, S.C., Molnar, F., Man-Son-Hing, M., Blair, R., Brosseau, L.,
Finestone, H.M., Lamothe, C., Komer-Bitensky, N., and Wilson, K.G.
Predictors of driving ability following stroke: a systematic review.
Topics in Stroke Rehabilitation, 79(17).
Matsui, H., Udaka, F., Miyoshi, T., Hara, N., Tamura, A., Oda, M., Kubori,
T., Nishinaka, K., and Kameyama, M. (2006). Frontal Assessment Battery
and Brain Perfusion Image in Parkinsons Disease. Journal of Geriatric
Psychiatry and Neurology, 79(1), 41-45.
McKenna, P., Jefferies, L., Dobson, A., and Frude, N. (2004). The use of a
cognitive battery to predict who will fail an on-road driving test.
British Journal of Clinical Psychology, 43, 325 336.
Niemeier, J.P., Marwitz, J.H., Lesher, K., and Walker, W.C. (2007). Gender
differences in executive functions following traumatic brain injury.
Neuropsychological Rehabilitation, 77(3), 293-313.
Paviour, D.C., Winterbum, D., Simmonds, S., Burgess, G., Wilkinson, L.,
Fox, N.C., Lees, A.J., and Jahanshahi, M. (2005). Can the frontal
assessment battery (FAB) differentiate bradykinetic rigid syndromes?
Relation of the FAB to formal neuropsychological testing. Neurocase,
Radford, K.A., Lincoln, N.B., and Murray-Leslie, C. (2004). Validation of the
stroke drivers screening assessment for people with traumatic brain injury.
Brain Injury, 18(8), 775-786.
Rapport, L.J., Hanks, R.A., and Bryer R.C. (2006). Barriers to Driving and
Community Integration After Traumatic Brain Injury. Journal of
Head Trauma Rehabilitation, 27(1), 34-44.
Reitan, R. (1986). Trail Making Test: Manual for Administration and Scoring.
Tucson, AZ: Neuropsychology Press.
Royall, D.R., Lauterbach, E.C., Cummings, J.L., Reeve, A., and Rummans,
T. A. (2002). Executive control function: A review of its promise and
challenges for clinical research: A report from the committee on
research of the American Neuropsychiatric Association. Journal of
Neuropsychiatry & Clinical Neurosciences, 14(A), 377-405
Schanke, A., and Sundet, K. (2000). Comprehensive driving assessment:
Neuropsychological testing and on-road evaluation of brain injured
patients. Scandinavian Journal of Psychology, 47,113-121.
Schultheis, M.T., Matheis, R.J., Nead, R., and Deluca, J. (2002). Driving
Behaviors Following Brain Injury: Self-Report and Motor Vehicle
Records. Journal of Head Trauma Rehabilitation, 17(1), 38-47.
Slachevsky, A., Villalpando, J.M., Sarazin, M., Hahn-Barma, V., Pillon, B.,
and Dubois, B. (2004). Frontal Assessment Battery and Differential
Diagnosis of Frontotemporal Dementia and Alzheimer Disease.
Archives of Neurology, 61, 1104-1107.
Soderstrom, S.T., Petterson, R.P., and Leppert, J. (2006). Prediction of driving
ability after stroke and the effect of behind-the-wheel training.
Scandinavian Journal of Psychology, 47, 419-429.
Stolwyk, R.J., Charlton, J.L., Triggs, T.J., Iansek, R., and Bradshaw, J.L.
(2006). Neuropsychological Function and Driving Ability in People
with Parkinsons Disease. Journal of Clinical and Experimental
Neuropsychology, 28, 898-913.
Tamietto, M., Torrini, G., Adenzato, M., Pietrapiana, P., Rago, R., and Perino,
C. (2006). To drive or not to drive (after TBI)? A review of the
literature and its implications for rehabilitation and future research.
Neurorehabilitation, 21, 81-92.
Yale, S.H., Hansotia, P., Knapp, D., and Ehrfurth, J. (2003). Neurologic
Conditions: Assessing Medical Fitness to Drive. Clinical Medicine
and Research, 7(3), 77-88.