Driving After Pediatric Traumatic Brain Injury: Impact of Distraction and Executive Functioning

Abstract

Objective:

The objective of the current study was to examine the driving performance of young drivers with a history of moderate to severe traumatic brain injury (TBI) compared with an uninjured control group. The impact of cell phone related distraction (conversation and texting) and executive functioning (EF) were also explored.

Method:

Individuals aged 16–25 years with (n = 19) and without (n = 19) a history of TBI engaged in a simulated drive under 3 distraction conditions (no distraction, cell phone conversation, and texting). Mean speed, maximum speed, standard deviation of speed, standard deviation of lane position, and crash rates were used as outcomes. The Global Executive Composite (GEC) from the Behavior Rating Inventory of Executive Functioning (BRIEF) was used to measure EF.

Results:

Significant Injury × Distraction × GEC interaction effects were noted on max speed and speed variability, with a trending Distraction × GEC interaction noted for lane position variability. The effect of distraction was most notable among individuals with greater GEC scores, across both injury groups.

Conclusions:

A history of pediatric TBI did not specifically impact driving performance independent of EF, with EF playing a central role in functioning across domains of driving performance. Consistent effect of EF suggests that deficits in driving performance may be associated with EF specifically, with individuals with EF difficulties following TBI at greater risk for poor driving performance.

Introduction

Pediatric traumatic brain injury (TBI) is the leading cause of acquired disability in childhood (Thurman, 2016). Survivors of pediatric TBI may have persistent impairments that delay their achievement of developmental milestones, such as learning to drive during adolescence (Babikian, Merkley, Savage, Giza, & Levin, 2015). Despite these potential delays, adolescents with TBI express the desire to drive like their peers (Di Battista, Soo, Catroppa, & Anderson, 2012). Further, executive functioning (EF) deficits and attention problems, skills critical to safe driving behavior, are common sequelae following pediatric TBI (Anderson, Brown, Newitt, & Hoile, 2011; Anderson, Godfrey, Rosenfeld, & Catroppa, 2012; Fay et al., 2009; Karver et al., 2012). Even injuries occurring at an early age can have a persistent impact with the CDC estimating that at least 5.2 million Americans have a lifelong or long-term need for support following TBI (Faul, Xu, Wald, & Coronado, 2010). Moreover, attention and EF deficits may emerge years following injury, particularly with the complex executive skills required to meet the emerging environmental demands associated with adolescence (Schwartz et al., 2003).

Given the long-term and emerging consequences of pediatric TBI, and documented influence of EF on driving performance, including crash frequency (Barkley, Murphy, Dupaul, & Bush, 2002), simulated driving performance (Mäntylä, Karlsson, & Marklund, 2009), and on-the-road driving performance (Demireva, McInerney, & Suhr, 2012; León-Domínguez, Solís-Marcos, Barrio-Álvarez, Barroso y Martín, & León-Carrión, 2017), adolescent and young adult drivers with a history of TBI may be at an increased risk for poor driving. Unfortunately, virtually no published research has examined the driving performance of novice drivers with a history of TBI, although there is a small but concerning literature reporting on the driving performance of adults following TBI. In the adult literature, the risk for driving-related impairments remains high even years after returning to driving (Bivona et al., 2012; Formisano et al., 2005; Pietrapiana et al., 2005). The neuropsychological predictors of driving problems after adult TBI are not entirely clear, but some have suggested that deficits in EF, attention, processing speed, working memory, and perceptual motor skills may be important (Tamietto et al., 2006). Notably, poorer EF is associated with on road driving performance (Marshall et al., 2007), poor risk awareness (Lundqvist et al., 1997), and increased likelihood of failing on-road exam (Devos et al., 2011; Hargrave, Nupp, & Erickson, 2012; Lundqvist et al., 1997; Schanke & Sundet, 2000) among adults with acquired brain injury. The potential impact of EF among adolescent/young adult and/or novice drivers is particularly noteworthy in light of evidence that executive dysfunction is one of the most common and persistent deficits following pediatric TBI (Anderson & Catroppa, 2005; Babikian & Asarnow, 2009; Mangeot, Armstrong, Colvin, Yeates, & Taylor, 2002; Narad et al., 2016; Sesma, Slomine, Ding, McCarthy, & Grp, 2008), with impairments emerging even years after injury (Muscara, Catroppa, & Anderson, 2008; Nadebaum, Anderson, & Catroppa, 2007; Sesma et al., 2008). Further, individuals with a distant injury history and general overall positive recovery may be at risk for emerging deficits in EF (Schwartz et al., 2003), making them a group at higher, albeit potentially unrecognized, risk for unsafe driving behaviors.

Although the adult literature helps to shed light on the impact of TBI on driving performance, it fails to address the increased risk of young/novice drivers. Specifically, motor vehicle crashes (MVCs) are the leading cause of death and injury among adolescents, (Centers for Disease Control and Prevention, 2010; National Highway Traffic Safety Administration, 2008) with rates of MVC significantly higher for young drivers compared with middle aged drivers (McKnight & McKnight, 2003). In fact, drivers aged 16–24 years old have the highest rates of MVCs (National Highway Traffic Safety Administration, 2017). Additionally, more novice drivers have higher rates of critical events (near crashes, rear-end crashes, road departure, side-swipe, and animal near crashes) compared with experienced adult drivers (Seacrist et al., 2016; Simons-Morton et al., 2015; Simons-Morton et al., 2011; Simons-Morton, Zhang, Jackson, & Albert, 2012).

A contributing factor to driving errors by adolescent and young adults is the increasing use of mobile electronics (e.g., cell phones) while driving. Engaging in phone tasks (e.g., talking, texting, wayfinding) results in behaviors such as looking away from the road, taking hands off the steering wheel, or directing attention away from driving. Such instances of distracted driving are cited as one of the primary causes of MVCs (Neyens & Boyle, 2008). Although many contextual factors (e.g., radio, passengers) contribute to distracted driving, cell phone–related, distracted driving fatalities are an increasing phenomenon and account for an estimated 18% of distraction-related driving deaths (National Highway Traffic Safety Administration, 2011). Although most drivers view distracted driving as a risk or threat to safety (AAA Foundation for Traffic Safety, 2016), 77% of drivers report engaging in cell phone conversations (Tison, Chaudhary, & Cosgrove, 2011), 81% of young adults write text messages while driving, and 92% read text messages while driving (Atchley, Atwood, & Boulton, 2011). Drivers aged 19–24 report the second highest rate of reading and typing texts and e-mails while driving (behind 25–39 year olds), and drivers aged 19–24 are most likely to report that typing or texting while driving is acceptable (AAA Foundation for Traffic Safety, 2016).

The detrimental impact of both hand-held and hands-free cell phone use on driving performance has been well documented (Beede & Kass, 2006; Kass, Cole, & Stanny, 2007), and Strayer and Johnston (2001) documented similar deficits in hands-free versus hand-held cell phone conversation. In simulated drives, individuals engaged in a cell phone conversation missed twice as many traffic signals (Strayer & Johnston, 2001), displayed decreased vehicular control (Drews, Pasupathi, & Strayer, 2008; Strayer & Johnston, 2001), and had significantly slower braking RTs compared with their performance with no distraction or while listening to the radio (Consiglio, Driscoll, Witte, & Berg, 2003). Distraction is particularly problematic for novice drivers, who are at a fourfold increased risk of being involved in MVCs resulting in injuries when talking on a cell phone (McEvoy et al., 2005). Further, a video analysis of teen driving performance estimated that distraction was a factor in nearly 60% of moderate to severe crashes, which is almost four times the estimates based on police reports (Carney, McGehee, Harland, Weiss, & Raby, 2015). Other estimates suggest that distraction was a factor in 58% of crashes, 89% of roadway-departures, and 76% of rear-end crashes (National Highway Traffic Safety Administration, 2008).

We examined the driving performance of adolescent and young adult drivers with a history of TBI compared with uninjured controls, during two different kinds of the most common driving distractions (cell phone conversation and texting) with the following primary aims: (a) examine the impact of pediatric TBI on simulated driving performance (main effect of injury group) and (b) examine whether the impact of cell phone related distraction differentially effects those with a history of TBI (Injury Group × Distraction Condition interaction). Additionally, given the known importance of EF in driving performance (Barkley et al., 2002; Mäntylä et al., 2009), we also explored how EF may moderate the effect of TBI, distraction, or their interaction on simulated driving performance.

The quality of simulated driving performance is conceptualized based on the ability to consistently maintain an appropriate speed of travel (mean speed, maximum speed, and speed variability) and consistent maintenance of vehicle control (lane position variability). Given the combined effects of inexperience of adolescent and young adult drivers and the neuropsychological consequences of TBI, we predicted that individuals with TBI would display poorer driving performance than those without TBI (primary aim 1). Additionally, we expect that engagement in a cell phone conversation or texting would impair the performance of all participants, with the greatest impairment occurring during texting as this requires cognitive, manual, and visual distraction. However, in light of the difficulties with divided attention and management of cognitive demands experienced by adolescents and young adults with a history of TBI (Babikian & Asarnow, 2009; Catroppa, Anderson, Morse, Haritou, & Rosenfeld, 2007), we hypothesize that driving distractions, including cell phone conversation and texting, would impair the driving performance of those with a history of TBI to a greater extent than the effect of distraction among drivers without a history of TBI (primary aim 2). Specifically, we predicted that the decrement in driving performance observed with adolescent and young adults with TBI engaged in phone calls or texting use would be significantly greater than that observed in drivers without TBI. Finally, in exploring the moderating effect of EF, as measured by parent- or self- behavior rating, on simulated driving performance we hypothesize that (a) those with greater EF deficits (across both TBI and uninjured control groups) would demonstrate more impaired driving than those with lower levels of EF deficit and (b) those with greater EF deficits in the TBI group would demonstrate the greatest impairment in driving performance.

Method

Participants

A total of 50 participants (25 TBI, 25 uninjured control) aged 16–25 years of age, all with a valid driver’s license, were enrolled in the study; data from 38 individuals (19 TBI, 19 uninjured control) were included in analyses. Participants were given the option to complete all study activities in a single visit or across two separate visits. One participant with TBI was lost to follow-up before completing the simulated drive, and simulator data from 11 participants (six uninjured controls and five with TBI) were lost because of technical difficulties, resulting in complete simulator data for 19 uninjured controls and 19 individuals with a history of TBI. Table 1 provides demographic information. Participants in the TBI group were recruited from a Level 1 Pediatric Trauma Center in Ohio, had a history of hospitalization for a moderate-severe TBI in the past, and were required to have been approved to return to driving by their medical provider prior to their participation in the study. All documented GCS scores represent postresuscitation GCS. A severe TBI was defined as a GCS score less than or equal to 8. Moderate TBI was defined as GCS 9–12, or a GCS of greater than 12 with accompanied abnormal clinical brain imaging. Individuals in the TBI group had an average GCS of 11.77 (SD = 4.66), and an average age at injury of 11.04 years (SD = 5.00). See Table 2 for individual injury details. Participants in the uninjured control group were matched to the TBI sample on age and sex. Given the well documented impairment in driving performance among individuals with Attention Deficit Hyperactivity Disorder (ADHD), combined with our focus on understanding how the sequelae of pediatric TBI may impact driving performance, the uninjured control group were required to have fewer than three total Diagnostic and Statistical Manual of Mental Disorders (fifth edition) symptoms of ADHD, as assessed with the Kiddie Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (KSADS-PL; Kaufman et al., 1997). No symptom threshold were required for the TBI group. All study procedures where approved by the Institutional Review Board.

Table 1.

Participant Demographic Information

Characteristic	TBI (n = 19)	Control (n = 19)	Total (N = 38)
Age in years, M (SD)	19.39 (1.83)	19.41 (1.92)	19.40 (1.85)
Sex, n (%) male	11 (58%)	13 (68%)	24 (63%)
Race, n (%) non-white	1 (5%)	5 (266/%)	6 (16%)
Months of driving experience, M (SD)	33.61 (21.75)	33.16 (24.25)	33.38 (22.75)
Cell phone use with driving (ever), n (%) yes	16 (84%)	13 (68%)	36 (76%)
KSADS–ADHD dx, n (%) yes	2 (11%)	0 (0%)	2 (5%)
History of ADHD-parent report, n (%) yes	4 (21%)	0 (0%)	4 (11%)
ADHD medication, n (%) yes	2 (11%)	0 (0%)	2 (5%)
Antidepressant medication, n (%) yes	0 (0%)	3 (16%)	3 (8%)
BRIEF version, n (%) parent report	9 (47%)	8 (42%)	17 (45%)
BRIEF–GEC, M (SD)	53.89 (9.66)	45.63 (8.15)	49.39 (9.65)

Note. KSADS = Kiddie Schedule for Affective Disorders and Schizophrenia; ADHD = attention deficit hyperactivity disorder; dx = diagnosis; History of ADHD = as reported by participant or parent; BRIEF = Behavior Rating Inventory of Executive Function; GEC = Global Executive Composite.

Table 2.

Injury Details for All TBI Participants

Participant	Age at injury	GCS	Accident details	Injury details
1	16 years, 2 months	4	Hit by car	Hemorrhagic contusions, extra-axial hematoma
2	11 years, 9 months	15	Fall while skiing	Subarachnoid hemorrhage
3	0 years, 3 months	N/A	Fall	Small extra-axial hemorrhage
4	7 years, 7 months	12	Object fell on patient	Subdural hematoma, multiple skull fracture
5	16 years, 2 months	15	Fall	Subdural hematoma, subarachnoid hemorrhage, skull fracture
6	12 years, 10 months	15	Fall from skateboard	Subarachnoid hemorrhage, hemorrhagic parenchymal contusion
7	15 years, 5 months	15	Fall	Subarachnoid hemorrhage
8	5 years, 8 months	N/A	Fall	Intrahemispheric minimal traumatic subarachnoid hemorrhage
9	14 years, 8 months	15	Hit with Baseball	Small hemorrhagic contusions, small areas of subarachnoid hemorrhage
10	14 years, 5 months	3	Hit with Baseball	Subdural hematoma
11	11 years, 6 months	15	Fall sledding	Depressed skull fracture and small extra-axial hemorrhage
12	0 years, 3 months	N/A	Fall	Depressed skull fracture, extra-axial hematomas and subarachnoid blood
13	7 years, 6 months	15	Hit with object	Subarachnoid hemorrhage
14	10 years, 7 months	7	Hit while playing football	Subdural hemorrhage
15	16 years, 3 months	3	Motor Vehicle Crash	Normal CT with > 1 hour loss of consciousness
16	15 years, 5 months	10	Fall from truck	Subdural hematoma
17	7 years, 2 months	12	Fall while playing basketball	Normal CT
18	13 years, 6 months	10	Wrestling	Normal CT with > 1 hour loss of consciousness
19	12 years, 9 months	8	Fall while playing baseball	Normal CT

Note. GCS = Post-resuscitation Glasgow Coma Scale; CT = computed tomography scan.

Executive Functioning

Parents provided ratings of their child’s EF on the Behavior Rating Inventory of Executive Functioning (BRIEF; Gioia, Isquith, Guy, & Kenworthy, 2000). In cases where the parent report was not available (ex. Participant >18 years of age or living outside the home), the BRIEF-A self-report form was completed (Guy, Isquith, & Gioia, 2004). Although correlations between self- and parent-reported BRIEF scores are moderate at best (Egan, Cohen, & Limbers, 2019; Walker & D’Amato, 2006; Wilson, Donders, & Nguyen, 2011), this procedure was used to maximize sample size and power. The BRIEF is a standardized rating scale on which raters report the frequency (never, sometimes, often) of behaviors reflective of the child’s, or their own, EF. The BRIEF is a well-validated measure of daily behaviors associated with EF that are often affected following TBI (Gioia & Isquith, 2004; Sesma et al., 2008). The General Executive Composite Scale (GEC) from the parent report BRIEF, or self-report BRIEF if parent report unavailable, was the measure of EF and used as the moderating variable, with greater GEC scores indicative of greater executive dysfunction. Parent-reported GEC was available for 45% of the sample (47% of TBI sample and 42% of control sample). It should be noted that rater (parent vs. self-report) did not significantly influence mean GEC scores (Parent: M = 49.06, SD = 9.67; Self-report: M = 49.68, SD = 9.89; t(34) = .19, p = .85), or likelihood of being in the high versus low GEC group, χ²(1) = 1.00, p = .32. Two participants (one TBI and one control) had GEC scores in the clinically elevated range (>65).

Procedure

During a baseline visit, all participants and their parents (when appropriate) provided informed consent. Parents and adolescents completed the KSADS and BRIEF. If individuals were on stimulant medication, they (and their parents where appropriate) were asked to report symptoms off medication as best they could. Parents (or participant if ≥18 years of age) completed a demographic form which gathered background information, injury information (for those in the TBI sample), and psychosocial and medical history. Adolescents also reported months of driving experience and whether they engaged in cell phone use while driving. Participants then completed the simulated drive. Participants taking stimulant medication refrained from taking medication the day of the simulated drive (see Table 1).

Driving Simulator

Participants completed a 40-min drive on a STISIM Model 400 simulator (Systems Technology INC). The system has three driving displays, a 135-degree driver field-of-view with integrated rear-view and side mirrors, a full-size steering wheel with dynamics-based feedback, full sized foot pedals, and an adjustable full-sized car seat. The roadway consisted of two lanes separated by a dashed yellow line and proceeded through urban and suburban settings. The drive consisted of sections of straight and curving roadways with other vehicles in the driver’s lane as well as the opposite lane of travel. The speed limit was 45 mph, and speed limit signs were posted along the roadway throughout the drive.

Prior to the start of the experimental drive, participants completed a 3-min practice drive to orient them to the simulator and simulator controls. Participants were then instructed to “drive as you normally would,” and they were told that during the drive they would receive telephone calls and text messages to which they needed to respond. Participants practiced using a text-enabled cell phone equipped with a hands-free headset. The first 10-min (40,000 ft) section of the experimental drive was an adjustment period during which participants became more comfortable with the driving simulator. The remaining 30 min (120,000 ft) were divided into three separate (but continuously driven) 10-min (40,000 ft) sections. During each period, participants were engaged in a hands-free cell phone conversation, a text message exchange, or no distraction. The order of the three conditions was randomized and counterbalanced across participants and each order of conditions occurred equally across groups.

During the conversation and texting conditions, an experimenter seated behind a room divider outside of view of the driver and simulator screen engaged the participant in either a cell phone conversation or continuous text message exchange. The content of the conversation and texting interactions were guided using two lists of randomly selected questions from The Book of Questions (Stock, 1985), with questions ranging from simple (i.e., what is your favorite food?) to more complex (i.e., if you found a wallet with $5,000, what would you do?). The use of the two lists was counterbalanced across the conversation and texting conditions. These lists were developed to keep participant engaged throughout the interaction, and specific responses to questions were not of interest.

During each of the three experimental conditions, one unexpected event occurred (car suddenly pulling in front of driver or a pedestrian suddenly crossing the street in front of the vehicle). Driving speed and lateral position were sampled every 30 ms during the entire drive. The first 4,000 feet (approximately 1 min) of each condition was removed from all analyses to control for carry-over effects across conditions. Also, because participants’ responses to unexpected events (braking and swerving) influenced measures of speed and lateral position, the 1,000 feet (approximately 15 s) following the unexpected event were removed from analyses of continuously measured variables (speed and lateral position). The remaining data were summarized by calculating the maximum, mean, and standard deviation (SD) of speed in mph, and SD of lateral position in feet for each section, resulting in two observations of each dependent variable per distraction condition. These variables were chosen because they demonstrate driving behavior reflective of the driver’s ability to control their vehicle on the roadway in terms of speed and vehicle position on the roadway and are sensitive to factors that may influence driving safety such as distraction or driver characteristics (i.e., age or clinical disorders). Speeding, as measured by mean and maximum speed (Aarts & van Schagen, 2006; Petridou & Moustaki, 2000) and speed variability (SD of speed; Lee, Saccomanno, & Hellinga, 2002; Narad et al., 2013; Shinar, Tractinsky, & Compton, 2005; Weafer, Camarillo, Fillmore, Milich, & Marczinski, 2008), are well understood and documented indicators of poor vehicle control and risk factors for motor vehicle crashes. Lane position variability (SD of lateral position) is a measure of within lane deviation, or weaving of the car, and represents driving precision or the driver’s ability to maintain the car within the lane of travel. This variable has been shown to be driver impairment (owing to intoxication, clinical disorder, etc.) (Arnedt, Wilde, Munt, & MacLean, 2001; Narad et al., 2013; Shinar et al., 2005; Weafer et al., 2008) and is reflective of vehicle control. Finally, if the participant’s vehicle made contact with the deployed object (car or pedestrian), a crash was coded.

Data Analysis

Separate 2 (group: control vs. TBI) × 3 (distraction: no distraction vs. conversation vs. texting) mixed-model analysis of variance were conducted for each of the dependent variables (mean speed, maximum speed, speed variability, and lane position variability). The moderating effect of EF was examined by including the Group × Distraction × GEC score into the models. Alpha levels less than .05 were interpreted as statistically significant in the above models. The continuous GEC score was used in all models, and in cases of significant effects of GEC, a median split (median = T score of 47.5) was used to understand the nature of the effects. Due to the artificial nature of this division, selected contrasts within significant interactions were interpreted as significant at the p ≤ .10 level. These selected contrasts include between injury groups/within distraction condition (i.e., TBI + low GEC vs. TBI + high GEC vs. Control + Low GEC vs. Control + Higher GEC in the conversation condition) as well as within injury group across distraction conditions (i.e., TBI + Low GEC during the conversation condition vs. TBI + Low GEC during the texting condition). Chi-square test was used to determine whether the rate of crashes differed by injury group and distraction condition.

Results

Mean Speed

No significant main effects or interactions were noted.

Maximum Speed

Analyses revealed a significant Injury Group × Distraction Condition, F(2, 172) = 3.13, p = .05, which can be best understood through the significant Injury Group × Distraction Condition × GEC Interaction, F(2, 172) = 3.67, p = .03; Figure 1. Interestingly, differences in maximum speed across conditions were observed among participants with higher GEC scores. Specifically, control participants with higher GEC scores had greater maximum speed during the no distraction condition compared with the conversation condition, t(172) = 3.06, p = .003, d = .98 and texting condition, t(172) = 1.91, p = .06, d = .61, and participants with a history of TBI and greater GEC scores demonstrated greater maximum speed during the conversation condition compared with the texting condition, t(172) = 1.76, p = .08, d = .40.

Figure 1.

The Injury × Distraction Condition × GEC interaction on maximum speed. GEC = Global Executive Composite; control low = control group with low GEC scores; control high = control group with GEC scores; TBI low = traumatic brain injury group with low GEC scores; TBI high = traumatic brain injury group with high GEC scores. Speed limit posted as 45 mph throughout the drive.

Speed Variability

A significant Injury Group × Distraction Condition was noted, F(2, 172) = 4.84, p = .01 which can be best understood through the Injury Group × Distraction Condition × GEC interaction, F(2, 172) = 4.76, p = .01, Figure 2. The effect of injury group and GEC were most notable within the conversation condition such that those with a history of TBI and higher GEC scores demonstrated greater speed variability than those in the control group with lower GEC scores, t(172) = 2.05, p = .04, d = .80, as well as those in the control group with higher GEC scores, t(172) = 1.67, p = .098, d = 1.11. The only distraction-related difference was noted for those in the control group with greater GEC scores, with greater speed variability noted during the texting condition compared with the conversation condition, t(172) = 1.89, p = .06, d = .70.

Figure 2.

The Injury Group × Distraction Condition × GEC score on speed variability. GEC = Global Executive Composite; control low = control group with low GEC scores; control high = control group with GEC scores; TBI low = traumatic brain injury group with low GEC scores; TBI high = traumatic brain injury group with high GEC scores.

Lane Position Variability

A main effect of GEC was noted, F(1, 172) = 4.36, p = .04, with individuals with greater GEC scores, regardless of injury group, demonstrating greater lane position variability (poorer vehicle control) across all conditions (low GEC: M = 1.09, SE = .08; high GEC: M = 1.21, SE = .07). The Distraction Condition × GEC interaction approached significance, F(2, 172) = 2.74, p = .07, Figure 3. Those with greater GEC scores demonstrated less lane position variability during conversation than during both the no distraction, t(172) = 3.16, p = .002, d = .78 and texting, t(172) = 4.70, p < .0001, d = 1.16 conditions. Similarly, those with lower GEC scores demonstrated greater lane position variability during the texting condition compared with the no distraction, t(172) = 1.69, p = .09, d = .44 and conversation condition, t(172) = 1.94, p = .05, d = .51. Of note, those with lower GEC scores displayed similar lane position variability during the no distraction and conversation conditions.

Figure 3.

The Distraction Condition × GEC interaction effect on lane position variability. Low GEC = low Global Executive Composite scores; high GEC = high Global Executive Composite scores.

Crash

There were not significant differences in the rate of crashes between the TBI and control group in any of the distraction conditions (no distraction: χ²(1) = 3.12, p = .08; conversation: χ ²(1) = 0.00, p = 1.00; texting: χ ²(1) = .21, p = .64).

Discussion

Surprisingly, a history of pediatric TBI did not significantly impact driving performance, independent of EF. Rather, EF, as measured by parent or self-report, played a central role in functioning across domains of driving performance, including maximum speed, speed variability, and lane position variability. EF directly affected lane position variability, a proxy for vehicle control, such that individuals with greater EF difficulty (although broadly within normal limits) demonstrated poorer vehicle control across all distraction conditions regardless of injury group. Interestingly, the EF by distraction condition effect approached significance and differential effects of distraction were noted depending on level of executive functioning. Those with greater EF difficulty demonstrated the lowest lane position variability during the conversation condition compared with the no distraction and texting conditions. In contrast, those with lower levels EF difficulty displayed the greatest lane position variability during the texting condition, compared with the no distraction and conversation conditions. EF also significantly moderated the effect of injury and distraction on maximum speed and speed variability. Although the effect of distraction was most noted among individuals with greater EF difficulty, the pattern/specifics of effects differed for those with and without a history of TBI. To our knowledge this is one of the first simulator studies that focuses exclusively on adolescents with a history of TBI, extending our knowledge of TBI-related driving impairments to adolescents. Further, findings highlight the role of EF as a behavioral/cognitive risk factor for impaired driving performance in adolescents.

Consistent with expectations and previous research (Drews, Yazdani, Godfrey, Cooper, & Strayer, 2009; Hosking, Young, & Regan, 2009; Narad et al., 2013; Neubauer, Matthews, & Saxby, 2012), distraction, particularly reading and sending text messages, appeared to have more pronounced impact on driving performance (i.e., maximum speed, speed variability, and lane position variability) for subgroups of our sample than voice communications. The need to divert one’s visual gaze from the road and hands from the steering wheel while texting creates a visual and manual distraction that impairs the ability to maintain central lane position. Interestingly, texting impacted driving performance, most notably for those with higher levels of EF difficulties, by slowing drivers down, creating more variability in speed and greater lane position variability. It has been suggested that texting while driving strains cognitive load because of the cognitive, visual, and manual aspects of the task, and individuals with lower levels of EF skills may have more difficulty managing the cognitive demands of this divided attention task. As a result, individuals may compensate by reducing speed (Crisler et al., 2008; Narad et al., 2013; Shinar et al., 2005); however, this is occurring within the context of greater speed variability within some subgroups (i.e., individuals in control group with greater executive functioning difficulties). Although slowing down may be beneficial in some driving situations, reductions in speed, particularly when occurring irregularly or in an unpredictable pattern, can impact traffic congestion (Cooper, Vladisavljevic, Medeiros-Ward, Martin, & Strayer, 2009) and highway safety (Aarts & van Schagen, 2006; Cirillo, 1968; Research Triangle Institute, 1970; Solomon, 1964).

In contrast, and similar to findings in Narad et al. (2013) and others (Engström, Johansson, & Ostlund, 2005; Shinar et al., 2005), engagement in a cell phone conversation was not as impairing. It is worth noting, however, that all participants used a hands-free headset during the conversation condition, which may have limited the manual/visual distraction associated with a cell phone conversation. Additionally, because the initial portion of each condition was removed from analyses to help limit any potential carryover effects from other conditions, it is likely that we failed to capture/detect the impact of handling and answering the phone call. Additionally, not only was engagement in a cell phone conversation while driving not impairing for at least one subgroup (individuals with greater levels of executive dysfunction), conversation improved driving performance with respect to lane position variability. These findings are consistent with Atchley and Chan (2011), who report decreased variability in lane position when engaged in a cell phone conversation during boring drive, suggesting that a concurrent cognitive task can improve performance during drives when vigilance is low. Participants in the present study completed a fairly lengthy simulated drive with no radio or source of entertainment, thus it is possible that the cell phone conversation served to decrease the monotony of the task compared with the no distraction condition. Finally, engagement in a verbal task while driving is associated with centralized eye gaze (Engström et al., 2005; Kingery et al., 2015; Nunes & Recarte, 2002; Recarte & Nunes, 2003), which may limit lane position variability. Although this may result in less lane position variability (often interpreted as improved vehicle control), it may come with a cost such as inattention blindness (diversion of attention from visual field to internal context involved in conversation) and an impaired ability to respond to peripheral events (Strayer, Drews, & Johnston, 2003). Although we didn’t find a significant increase in crashes during the conversation condition, all events occurred in the center of the individual’s visual field. Had these events occurred more peripherally, the negative effects of conversation may have been more evident.

Interestingly, distraction did not impact the driving performance of individuals in the uninjured control group with lower levels of executive dysfunction; however, distraction had a clear effect of those in the uninjured control group with more difficulties with EF, again highlighting the importance of EF in driving performance. Because maintaining a consistent speed and a stable, central lane position requires constant attention to the road and one’s surroundings, the moderating effect of EF is not surprising (Rakauskas, Gugerty, & Ward, 2004).

These findings should be viewed in the context of the limitations. First, driving performance was examined in the context of a simulator. Although it is an artificial driving environment that only captures a sampling of driving behavior, studies have cited the validity of simulator use, noting that it is a safe and controlled method for assessing high-risk driving behaviors (Blaauw, 1982; Lee, Lee, & Cameron, 2003; Lew et al., 2005). Specifically, relative performance within a simulator (maintenance of speed and control/direction) is highly correlated with on the road driving performance (Godley, Triggs, & Fildes, 2002; Harns, 1996; Shechtman, Classen, Awadzi, & Mann, 2009; Tornros, 1998) and even predictive of on-road driving performance (de Winter et al., 2009; Lew et al., 2005). Indeed, it would be very difficult to ethically induce similar driving risks (e.g., sudden entry of a vehicle onto the roadway) outside of a simulator. Further, because of the nature of the simulator, and the vast range of variables collected by the simulator, there are no norms of performance on this task. Although this precludes us from being able to operationalize the boundaries for safe or normal driving performance based on simulator outcomes, it provides us with valuable information regarding the patterns of performance across groups. Future research would benefit from including official driving records or in car monitoring devices to have a clearer picture of real world driving behaviors outside of the simulator. Another limitation is that the driving scenario included only suburban and urban driving roadways. The work of Reimer and colleagues (Reimer, Mehler, D’Ambrosio, & Fried, 2010) suggests that roadway factors may influence driving outcomes, and the effect of distraction may vary by environment. The impact of distraction on different roadway types (e.g., highway settings) and conditions (e.g., adverse weather and traffic) should be examined in future studies to further understand the impact of distracted driving on young drivers. Also, with regard to the sample, the uninjured control group was limited to those with limited symptoms of ADHD, and generally speaking both groups had few comorbidities and EF largely within the normal range, which may not be representative of broader adolescent population or the broader population of adolescent survivors of pediatric TBI. It is worth noting that three individuals in the control group were being treated with antidepressant medication for depression and/or anxiety, and although the EF skills of these individuals did not vary greatly from the larger sample, the small number of individuals in this subgroup precluded examination of the impact of these comorbidities. Some studies have suggested that certain comorbidities (e.g., oppositional defiant disorder/conduct disorder) increase driving risk (Thompson, Molina, Pelham, & Gnagy, 2007), and future studies including a broader range of comorbidities would allow for a better understanding of adolescent driving risk factors. We also included individuals who were already cleared to return to driving, which likely eliminated individuals with more severe TBI for whom caregivers may have concerns about their abilities to drive safely, or the adolescents ability to pass the driving exam, and may also have limited their motivation to perform well on the simulated driving tasks. Future work would be improved by including adolescents at varying points in TBI recovery to better understand driving performance throughout recovery. Additionally, our sample size was small, with the study sufficiently powered to examine the main effects of TBI and distraction as well as their interaction on driving performance; however, the moderation analyses were likely underpowered. This, coupled with the loss of simulator data from 12 participants, potentially limited our ability to detect group differences. Finally, EF is a complex multifaceted construct and our study utilized a parent-reported or self-reported behavioral ratings of EF, which may not be highly correlated with performance or laboratory measures of EF. Given the present findings and clear importance of EF in driving performance, future studies should include performance-based measures of EF to better understand what aspects of EF are strongly related to driving performance.

The consistent effect of EF, despite participants having scores largely within the normal range, suggests that deficits in driving performance may be associated with EF specifically, with individuals with EF difficulties following TBI at greater risk for poor driving performance. It is true that our sample utilizes a generally high functioning group of TBI survivors with arguably normal EF skills; however, this further highlights the notion that adolescents and young adults with a history of TBI, even a distant history, who appear to have essentially recovered should be monitored for EF problems throughout development to mitigate the onset of impairments in tasks that present themselves throughout development. The central role of EF on the driving performance of adolescents with and without a history of TBI suggests that driving interventions targeting adolescents with EF deficits (regardless of clinical population) are required. Although we did not find that youth with a history of TBI were uniquely vulnerable to the effects of distraction while driving, this is of dubious comfort; texting worsened driving across all groups. Global effects of texting on driving performance supports public policy and education surrounding the risks associated with texting while driving. Providers should play a central role in disseminating this safety-related information, and parents may benefit from additional support to set and uphold boundaries surrounding texting while driving.

Impact and Implications.

This study examines the driving performance of young drivers with a history of traumatic brain injury (TBI), as well as the impact of cell phone–related distraction and executive functioning. The pattern of findings suggests that executive functioning, rather than TBI specifically, played a central role in functioning across domains of driving performance. Findings highlight that adolescents and young adults with a history of TBI, even a distant history, should be monitored for executive functioning deficits throughout development to help mitigate the onset of new challenges as they engage in new/complex tasks.