Associations Between Post-Traumatic Stress Disorder Symptoms and Automobile Driving Behaviors: A Review of the Literature

Abstract

Human factors are responsible for most motor vehicle accidents that occur on the road. Recent work suggests that symptoms of posttraumatic stress disorder (PTSD) are linked to reduced driving safety, yet none have provided a comprehensive review of this small, emerging literature. The present review identified twenty-two studies reporting associations between PTSD and driving behaviors. Among these, longitudinal designs (k=3) and studies using objective driving performance measures (e.g., simulators) (k=2) were rare. Most studies (k=18) relied on brief screener measures of PTSD status/symptoms or a prior chart diagnosis, while few used a standardized structured interview measure to determine PTSD status (k=4), and only a small number of studies assessed PTSD symptom clusters (k=7). PTSD was most frequently associated with increased rates of hostile driving behaviors (e.g., cutting off others), unintentional driving errors (e.g., lapses in attention) and negative thoughts and emotions experienced behind the wheel. Findings regarding risk of motor vehicle accident and driving-related legal issues were variable, however relatively few studies (k=5) explored these constructs. Future directions are discussed, including the need for work focused on concurrent PTSD symptom/driving-related changes, more comprehensive PTSD and driving assessment, and consideration of the contributions of comorbid traumatic brain injury history and other neurological and psychiatric conditions on driving outcomes.

1. Introduction

Driving is a vital component of daily living and is essential to one’s sense of autonomy (Burkhardt, 1999; Ragland, Satariano, & MacLeod, 2005). However, it is also a leading source of death and disability, as 1.2 million individuals die and 50 million are injured in motor vehicle accidents each year (World Health Organization, 2015). Given that operating a vehicle requires the integration of numerous complex skills and abilities (Groeger, 2013), it is not surprising that human factors are considered the strongest predictors of accident risk (Petridou & Moustaki, 2000).

Approximately 7% of the population in the United States meets criteria for Posttraumatic Stress Disorder (PTSD) during their lifetime, making it one of the most highly prevalent psychiatric conditions (Kessler et al., 2005). PTSD is marked by intrusive symptoms related to the traumatic event (e.g., upsetting memories, nightmares), avoidance of trauma-related stimuli, negative alterations in cognitions or mood (e.g., inability to recall key features of trauma, feelings of isolation, difficulties experiencing positive emotions) and alterations in arousal and reactivity (e.g., difficulty concentrating or sleeping, irritability). Several of the symptoms most frequently observed in PTSD have been hypothesized to impact driving safety and behaviors (Lew et al., 2010). For example, the feelings of anger and irritability as well as risk-taking and hypervigilance behaviors characteristic of PTSD may carry over to behaviors and emotions behind the wheel, such as driving in an aggressive manner or evasive manner, or experiencing “road rage” (Hannold et al., 2013; Possis et al., 2014). PTSD may also be accompanied by disruptions in cognitive function, including attention, memory and executive function (DeGutis et al., 2015; Vasterling et al., 2009). Dysfunction in these same domains has been tied to accident risk and dangerous driving behaviors in other clinical populations (Asimakopulos et al., 2012; Colonna et al., 2016; Walker & Trick, 2018), raising concern that they may also undermine driving safety in those with PTSD. Even mere exposure to a traumatic event, particularly in the case of military Veterans, may convey a sense of invulnerability and thus impact risk-taking behaviors while driving (Kang & Bullman, 1996; Lew et al., 2010), although it should be noted that other explanations for the trauma exposure/risk-taking behavior relationship also exist (Ben-Zur & Zeidner, 2009). Overall, significant overlap between the multiple, interdependent capacities necessary to safely drive and the effects that symptoms of PTSD may exert on those same capacities suggests possible concerns about driving safety.

Prior literature reviews have only tangentially touched on the link between PTSD symptoms and driving behaviors. These reviews have generally described the broader work regarding driving behaviors and difficulties, specifically in Veteran cohorts, rather than providing a more comprehensive and focused discussion of the PTSD/driving relationship (Lew et al., 2010; Possis et al., 2014). In their review of Veterans’ driving behaviors, Lew and colleagues (2010) were among the earliest to propose that PTSD symptoms may negatively influence driving safety. However, these authors noted at the time that this topic was “largely unexamined” from an empirical standpoint (Lew et al., 2010). Similarly, Possis and colleagues (2014) summarized literature demonstrating that Veterans’ psychological symptoms following deployment (e.g., fear, anger, aggression, thrill-seeking) might undermine driving safety. While these symptoms are characteristic of PTSD, they are neither exclusive to the disorder nor are they representative of all symptoms a patient with PTSD is likely to experience. This relative absence of prior PTSD/driving literature reviews stands in sharp contrast to the multitude of reviews on driving behavior written with respect to other psychiatric and neurological conditions, such as traumatic brain injury (TBI), Alzheimer’s Disease, anxiety and attention-deficit hyperactivity disorder (Classen et al., 2009; Hird et al., 2016; Jerome et al., 2006; Taylor, Deane, & Podd, 2008). A comprehensive review of the literature is necessary given the recent proliferation of studies specifically exploring the association between PTSD and driving, the relatively high prevalence of PTSD in both the civilian and military populations, and the broader public health need to screen for human factors that may endanger other road users.

The overall goal of the present manuscript was to review and synthesize available literature pertaining to the effects of PTSD on driving behaviors. In particular, this review had two aims. The first aim was to summarize and critically review the various study methodologies employed in this literature, including designs, samples and measures used. Critiques of these methods are outlined in the Discussion section of this manuscript. The second aim was to determine the specific facets of PTSD and particular aspects of driving that are most strongly and consistently associated with one another, with a specific focus on aspects of driving that may endanger an individual with PTSD and/or their fellow road users.

2. Material and Methods

2.1. Search Strategy

Electronic databases PubMed and PsychInfo were searched from inception to August 20 2021. This search was conducted by the first author (JPKB, a doctoral-level clinical psychologist) and confirmed independently by the other three authors (all doctoral-level psychologists).

In order to identify as many pertinent studies as possible, a broad search strategy using the keywords “PTSD” OR “post-traumatic stress disorder” OR “trauma” combined with the keywords “driving” OR “driver” OR “automobile” was employed. The abstracts for all resulting citations were screened by the first author (JPKB) to assess eligibility for inclusion in this review. If the contents of the abstract were insufficient to determine eligibility, the full text was screened. Reference lists for all studies included in the review were screened to determine if any studies not identified in the database search warranted inclusion.

2.2. Inclusion/Exclusion Criteria

For a manuscript to be included in this review, the study had to: (1) Include a measure of the presence/absence of a current PTSD diagnosis or a measure of current PTSD symptom severity; (2) Include a measure of current or recent past (i.e., within the past month) driving behaviors; and (3) Have conducted analyses examining the relationship between the measure of PTSD and the measure of driving behaviors. The Criterion A trauma was required to have occurred prior to the period on which the participants’ driving behaviors were assessed to maintain a focus on the effects of PTSD on driving. The PTSD measure used was required to be a structured interview, screener or self-report measure specifically targeting PTSD symptoms (e.g., PTSD Check List [PCL]) or recent chart review. Participants were not required to have received a formal diagnosis of PTSD in order to maximize the number of included studies as well as to assess the relationship between continuously measured PTSD symptoms and driving outcome variables. The driving measure had to directly assess driving behaviors, which were broadly interpreted to include driving maneuvers (whether intentional or accidental), emotions, accidents, status and frequency (see Data Analysis). As such, studies that solely assessed skills indirectly related to driving (e.g., cognitive tests, self-report measures of impulsivity or risk-taking without items specifically referencing driving) were excluded. The driving measure used was allowed to be subjective (e.g., self-report, structured interview) or objective (e.g., simulator, driving sensors). Studies that were reviews, non-English studies or duplicate studies were excluded.

Given the limited literature and high comorbidity of PTSD with other conditions (e.g., depression), participants were not required to be free from other psychiatric or neurological disorders. History of traumatic brain injury (TBI) is particularly frequently comorbid with PTSD, may interact with PTSD to impact functional outcomes, and has been the primary population of interest in several prior PTSD/driving studies (Bernstein et al., 2021; Lew et al., 2010; McGlinchey et al., 2017). As such, for those studies that met inclusionary criteria to be listed in the current review, prevalence and contributions of comorbid TBI to driving outcomes were also discussed.

2.3. Data Analysis

Given the broad search strategy, it was expected that many types of study designs would be represented in the studies identified. Since this heterogeneity would make quantitative analysis of these studies difficult, it was planned to instead provide a narrative analysis whereby studies were grouped and compared based on the measures used, methods employed, and outcomes produced.

For purposes of grouping studies based on related finding, studies were categorized as studying one or more of the following five outcomes: (1) Hostile/Aggressive Driving, or intentional behaviors that may have endangered the safety of other road users; (2) Driving Errors, or unintentional behaviors that may or may not have endangered the safety of other road users; (3) Driving Cognitions and Emotions, or thoughts and feelings an individual experienced when driving; (4) Accidents and Legal Issues; and (5) Driving Status, or whether or not an individual remained an independent driver. All studies examined at least one of these driving-related outcomes, and several studies examined more than one outcome.

3. Results

Using the search strategy described above, a total of 2,981 studies were identified. Of these, 22 manuscripts met the inclusion criteria and were included in the review. All studies were published between 2003 and 2021. Table 1 displays a summary of each of the studies included in the review, while Table 2 shows a summary of the driving domains assessed within each study.

Table 1.

Studies Included in Narrative Review (k=22)

Authors (year), country	Purpose	Design	Sample	PTSD Measure	Driving Measure	PTSD/Driving-Related Findings
Alavi et al (2017), Iran	To examine effects of personality, driving behaviors and mental health diagnoses on road traffic accident frequency	Cross-sectional	n=800 bus and truck drivers; recruited from referrals to hospital as prerequisite for renewing license	Psychiatric interview with binary (yes/no) measure of PTSD	Number of lifetime self-reported accidents	After accounting for other psychiatric variables and self-reported driving behaviors, PTSD diagnosis did not predict number of lifetime accidents
Amick et al (2013), USA	To evaluate driving simulator performance in OEF/OIF Veterans who self-report poor driving safety following deployment	Cross-sectional case-control	n=25 veterans and 25 civilian controls without psychiatric history; recruited from VA informational flyers and referrals from other research studies	PCL-Military	Simulator (Virtual RX Driver NDX System)	A non-significant trend indicated that greater PTSD symptom severity was associated with greater number of driving errors
Baker et al (2014), USA	To investigate whether a novel self-report measure of driving behavioral stress response, the Driving Behavior Survey (DBS), was sensitive to changes associated with PTSD treatment	Case-control follow-up	n=40 adults with MVA-related PTSD; recruited from community using flyers and public service announcements	CAPS	DBS (assesses anxiety-based performance deficits, hostile aggressive behaviors, and excessive safety/caution behaviors)	Greater PTSD symptoms were mildly to moderately associated with total and subscale DBS scores
Bernstein et al (2021), USA	To identify predictors of driving status in service members and veterans 1-year post-TBI	Prospec tive longitu dinal	n=471 service members and Veterans; recruited from VA TBI inpatient rehab clinics	PCL-C	Single item assessing current driving status (yes/no)	Higher PTSD symptom severity (particularly dysphoria) was associated with lower likelihood of remaining an active driver at 1-year postinjury
Caplan et al (2017), USA	To assess community reintegration problems among Veterans and military service members with mild or moderate/severe traumatic brain injury (TBI) one year following injury and identify unique predictors contributing to these problems	Prospec tive longitu dinal	n=154 veterans; recruited from VA TBI inpatient rehab clinics	PCL-C	Single self-report item assessing current driving status	Among those with mTBI, greater PTSD symptoms associated with lower likelihood of remaining an active driver
Classen et al (2011), USA	To examine differences in rates and types of driving errors between Veterans with comorbid mTBI history and PTSD and healthy controls	Cross-sectional case-control	n=18 vets with PTSD and mTBI history and 20 healthy controls; recruited from convenience sample from a prior study	PTSD in chart	Simulator (STISM M500W)	Veterans made more driving errors than controls, particularly overspeeding and adjustment to stimuli
Clapp et al (2014), USA	To evaluate the psychometric properties of the DBS among individuals with posttraumatic stress disorder (PTSD) following MVA	Cross-sectional	n=40 adults with MVA-related PTSD	CAPS	DBS	Higher PTSD symptom severity was associated with greater hostile/aggressive driving behaviors
Clapp et al (2019), USA	To examine characteristics of the DBQ in Veterans with PTSD	Cross-sectional	n=160 male veterans; recruited from convenience sample from prior PTSD intervention study	CAPS	DBS	Higher PTSD symptom severity associated with greater hostile/aggressive driving behaviors, anxiety-based performance deficits and exaggerated safety/caution behaviors
Fear et al (2008), UK	To assess whether deployment and deployment-related factors are associated with increased risky driving	Cross-sectional	n=8,127 UK veterans; recruited from a larger study of UK veterans who served in the 2003 Iraq War	PCL-C	Two self-report items, one each assessing frequency of speeding and seatbelt use	Greater PTSD symptom severity associated with increased risk of speeding and/or not wearing setbelt
Hannold et al (2013), USA	To explore driving perceptions of Veterans with comorbid PTSD and history of mTBI	Cross-sectional	n=5 Veterans with PTSD and mTBI; recruited from convenience sample from Classen 2011 study	PTSD in chart	Structured interviews on triggers for dangerous driving behaviors	While Veterans recognized environmental triggers of their dangerous driving and used strategies to improve their driving safety, they continued to drive aggressively
Hoggatt et al (2015), USA	To assess the prevalence of risky driving and its demographic, mental health, and deployment-related correlates	Cross-sectional	n=2,616 national guard members; recruited from study monitoring Ohio Army National Guard members’ mental health	PCL-C	Six self-report items, one each assessing frequency of: seat belt use; drinking and driving; passed driver on right; crossed intersection in red light; ignore speed limits at night or early morning; and underestimation of speed of oncoming vehicle	Relative to no mental health conditions, greater lifetime history of PTSD associated with higher prevalence of most risky driving behaviors assessed
Kuhn et al (2010), USA	To determine whether overall PTSD symptom severity and which symptom clusters are associated with risky driving	Cross-sectional	n=474 male veterans; recruited from VA residential PTSD treatment program receiving VA residential treatment for PTSD	PCL-M	Six self-report items, one each assessing frequency of: verbal outbursts/angry hand gestures while driving; tailgating intentionally cutoff, or chase other drivers; driving after drinking or using psychoactive drugs; driven vehicle into another object (e.g., another car, tree); overall aggressive driving; and use seatbelts	Two thirds of the sample reported lifetime aggressive driving and one third reported current aggressive driving; PTSD symptom severity was associated with aggressive driving, but not other types of unsafe driving
Lew et al (2011), USA	To examine associations between self-reported driving difficulties and TBI and PTSD	Cross-sectional case-control	n=205 veterans; recruited from VA polytrauma outpatient clinics	PTSD in chart	Brief Driving Questionnaire, which includes seven items assessing frequency of: general driving difficulties; warnings by the police for traffic violations; receipt of a traffic citation (ticket or fine); driving accidents; near misses while driving; problems with anger or impatience while driving, and becoming lost/disoriented while driving	Respondents who had a diagnosis of PTSD (with or without TBI) reported more severe driving difficulties since returning from deployment than did those with TBI but no PTSD and those with neither PTSD nor TBI history
Mairean et al (2019), Romania	To evaluate associations between driving cognitions, rumination, and PTSD symptoms	Cross-sectional	n=633 Romanian drivers involved in a MVA 3–12 months ago; secondary analysis of larger study of driving thoughts and behaviors	PCL	Driving Cognitions Questionnaire (20-item self-report measure assessing panic, accident and social-related driving cognitions)	PTSD associated with panic, accident and social-related driving cognitions
Mairean et al (2020), Romania	To explore relations between PTSD symptoms, fear and avoidance of driving, and aberrant driving behaviors	Cross-sectional	n=162 Romanian drivers involved in a MVA in the past 2 years; secondary analysis of larger study of driving thoughts and behaviors	PCL	Driving Behavior Questionnaire (24-item self-report measure of aberrant driving behaviors) and Travel Phobia Questionnaire (12-item measure of fear and avoidance of driving)	Greater PTSD symptoms associated with higher frequency of lapses in attention and accidental driving errors, but not intentional dangerous driving behaviors, and these relationships were found for men but not women
Saberi et al (2013), Iran	To examine demographic and occupational correlates with PTSD in Iranian commercial motor vehicle drivers	Cross-sectional case-control	n=424 commercial drivers; recruited randomly from a hospital driving training course	PCL-C	Self-reported number of accidents lifetime	Individuals who reached cutoff for suspected PTSD reported more lifetime accidents
Sayer et al (2010), USA	To describe prevalence and types of community reintegration problems faced by Iraq-Afghanistan combat veterans, measure their interest in interventions or information to promote readjustment, and assess associations between PTSD, reintegration problems and treatment interests	Cross-sectional case-control	n=1,226 veterans; recruited via mailed surveys to veterans enrolled in VA healthcare and meeting study inclusion criteria	Primary Care PTSD screen (PC-PTSD)	Single self-report item assessing whether others have noticed that participant engages in dangerous driving (yes/no)	Compared to those who denied the driving item, those who endorsed the item were four times as likely to screen positive for PTSD
Strom et al (2012), USA	To examine the relation between PTSD symptom severity and risk-taking behaviors	Cross-sectional	n=395 veterans; recruited from outpatient mental health and primary care services at a large VA medical center	PCL-Military	Risk-taking survey which included seven driving-related items assessing thrill-seeking, substance abuse and aggression	Greater PTSD symptom severity associated with more frequent risky driving behaviors
Tessier et al (2017), USA	To assess whether veterans with comorbid PTSD and TBI use coping self-instruction strategies when behind the wheel	Cross-sectional case-control	n=23 veterans w comorbid PTSD and mTBI, and 9 healthy veteran controls; recruited from larger study of post-9/11 Veterans’ driving behaviors	PTSD in chart	Drivers Angry Thoughts Questionnaire (65-item self-report measure of frequency of angry thoughts experienced while driving)	Relative to controls, Veterans with PTSD or TBI engaged in fewer coping self-instruction thoughts behind the wheel, despite experiencing greater anger
Van Voorhees et al (2018), USA	To examine relative contribution of PTSD and TBI to aggressive driving in a sample of combat veterans	Cross-sectional case-control	n=1,102 veterans in mid-Atlantic; recruited from larger study of post-9/11 Veterans eligible for VA care	PCL	Single self-report item measuring road rage/aggressive driving in past year (yes/no)	Controlling for relevant demographic variables, PTSD without TBI and PTSD with co-occurring TBI were associated with increased risk of road rage, but TBI without PTSD was not.
Whipple et al (2016), USA	To assess Veterans’ driving behaviors, driving-related anxiety, and driving-related emotional experiences	Cross-sectional case-control	n=23 combat Veterans diagnosed with comorbid TBI and PTSD and 10 healthy Veteran controls; recruited from polytrauma and primary care clinics at large VA	PTSD in chart	Veteran Driving Questionnaire (self-report measure of situational anxiety, driving behaviors and affective states)	Drivers with TBI and PTSD reported more frequent risky driving behaviors and higher levels of anxiety while driving than nondisabled combat Veterans.
Zinzow et al (2013), USA	To examine the role of driving-related anxiety, including common cues and mental health correlates, with risky driving behaviors	Cross-sectional case-control	n=46 recently deployed male service members; recruited by email through local and state service members and veterans’ organizations	PC-PTSD	Structured interviews evaluating prevalence of driving-related anxiety or hyperarousal in relation to combat exposure, PTSD, depression and help-seeking behaviors	Compared to those who screened negative for PTSD, Veterans who screened positive more frequently reported driving differently following deployment and reported experiencing greater anxiety/hyperarousal while driving

Notes: DBS = Driving Behavior Scale; MVA = Motor vehicle Accident; PC-PTSD = Primary Care PTSD Screen; PTSD = Postraumatic Stress Disorder; PCL = PTSD Checklist; TBI = Traumatic Brain Injury

Table 2.

Driving Domains Assessed in Relation to PTSD (k=22)

Authors (year), country	Hostile/Aggressive Driving Behaviors	Driving Errors	Cognitions and Emotions	Accidents and Legal Issues	Driving Status
Alavi et al (2017), Iran				−
Amick et al (2013), USA		+
Baker et al (2014), USA	+	+
Bernstein et al (2021), USA					+
Caplan et al (2017), USA					+
Classen et al (2011), USA	+	+
Clapp et al (2014), USA	+
Clapp et al (2019), USA	+	+
Fear et al (2008), UK	+
Hannold et al (2013), USA	+		+
Hoggatt et al (2015), USA	+			+
Kuhn et al (2010), USA	+			−
Lew et al (2011), USA		+	+
Mairean et al (2019), Romania			+
Mairean et al (2020), Romania		+
Saberi et al (2013), Iran				+
Sayer et al (2010), USA	+
Strom et al (2012), USA	+
Tessier et al (2017), USA			+
Van Voorhees et al (2018), USA	+
Whipple et al (2016), USA			+
Zinzow et al (2013), USA			+

Notes: “+” indicates a relationship was identified between PTSD symptom severity and the driving domain assessed, while “−” indicates no such relationship was identified

3.1. Study Methodologies

The first goal of this review was to summarize study methodologies (including designs, samples and measures, used in this literature. Of the studies reviewed, most (k=19) were cross-sectional in design (see Design column in Table 1). Of the remaining three studies, Baker and colleagues (2014) analyzed data at baseline, 6 weeks and 18-weeks following enrollment in a PTSD treatment setting, while Bernstein and colleagues (2021) and Caplan and colleagues (2017) performed secondary analyses of a TBI research database at baseline and approximately one year post-injury. Studies varied considerably in sample size, ranging from n=5 (Hannold et al., 2013) to n=8,127 (Fear et al., 2008). A majority of studies (k=16) recruited an exclusively Veteran/active service member sample or used a Veteran/active service member cohort as a case comparison group (see Sample column in Table 1).

3.1.1. PTSD Assessment

Half of the studies (k=11) used a self-report measure of PTSD symptom severity, specifically the PCL (see PTSD Measure column in Table 1). Four studies used a structured or semi-structured interview, with three using the CAPS and one using the Schedule for Affective Disorders and Schizophrenia. The remaining eight studies used a clinician-administered PTSD screener measure (i.e., PC-PTSD) (k=2) or a pre-existing chart diagnosis (k=6). Of those studies in which PTSD symptom cluster scores were available (i.e., PCL or CAPS; k=14), seven studies reported associations between PTSD symptom clusters and driving variables, while the other seven studies solely reported PTSD measure total scores or did not report these scores at all (i.e., participants met a cutoff score, but raw scores were not reported).

3.1.2. Contributions of TBI To Driving Variables

Seven studies included history of TBI among its inclusion criteria. Of these, only one (Lew, Patricia Woods, & Cifu, 2011) examined the independent contributions of TBI to driving variables. Regardless of TBI history, those with PTSD endorsed significant driving difficulties, whereas those with solely TBI history reported difficulties similar to those without either diagnosis. The remaining six studies in this subset did not delineate contributions of TBI from PTSD to driving. One additional study (Van Voorhees et al., 2018) did not require participants to have a history of TBI but did assess independent contributions of TBI history to driving variables. This study found that whereas PTSD with or without TBI history was associated with increased risk of road rage, TBI without PTSD was not.

3.1.3. Driving Assessment

Only two studies used an objective, performance-based measure of driving (i.e., a driving simulator), while the remaining 20 studies used a subjective measure (i.e., self-report or structured interview) (see Driving Measure column of Table 1). Specific measures selected varied widely between studies both in the breadth and depth of driving behaviors assessed. Whereas seven studies employed a validated driving measure used in previous studies, two created measures specifically for the study in question. An additional three studies employed individual driving-related items from measures tapping broader constructs (e.g., risk-taking) and six used single dichotomous (yes/no) measures of a particular behavior of interest (e.g., road rage/aggressive driving, motor vehicle accidents; driving status). Two studies used a structured interview measure, one to assess driving anxiety and related difficulties (Zinzow et al., 2013) and the other to assess environmental triggers, coping strategies and driving behaviors (Hannold et al., 2013).

3.2. Associations Between PTSD and Driving

The second aim of this review was to determine the specific facets of PTSD and particular aspects of driving that were most strongly and consistently associated with one another. In this review, half of the studies (k=11) explicitly noted that a goal was to examine the association between PTSD and any facet of driving (see Purpose column of Table 1). The other half (k=11) adopted a broader scope, with some (k=6) aiming to elucidate the effects of a broader or different range of mental health symptoms/conditions (e.g., TBI) and simply including PTSD as a covariate, others (k=4) exploring an outcome variable more wide-ranging than driving behaviors (e.g., risk-taking), or both (k=1) (Caplan et al., 2017). However, these latter eleven studies included examinations of associations between PTSD and at least one driving variable and thus were eligible for the current review.

3.2.1. Hostile/Aggressive Driving

A total of eleven studies identified associations between greater PTSD symptom severity and more frequent intentionally hostile or aggressive driving behaviors (see PTSD/Driving Findings column of Table 1). In their work using the Driving Behavior Scale (DBS), a study by Baker and colleagues (2014) and two studies by Clapp and colleagues (2014; 2019) found that greater PTSD symptom severity was broadly associated with more frequent hostile and aggressive driving behaviors. Kuhn and colleagues (2010) and Strom and colleagues (2012) found that greater PTSD symptom severity was associated with more frequent verbal outbursts or angry hand gestures at others while driving, tailgating, intentionally cutting off, chasing other drivers, and/or intentionally driving a vehicle into another object. Similar findings were reported by Van Voorhees and colleagues (2018), who found that individuals with PTSD were more likely to endorse driving aggressively and experience road rage. Fear and colleagues (2008) and Classen and colleagues (2011) both noted that greater PTSD symptom severity was linked to higher rates of self-reported speeding. In a large study of national guard members, Hoggatt and colleagues showed that greater PTSD symptom severity was associated with higher rates of passing drivers on the right, crossing an intersection during a red light, ignoring speed limits early or late in the day, and underestimation of the speed of an oncoming vehicle. Hannold and colleague (2013) found that Veterans with comorbid PTSD and mTBI reported driving aggressively despite incorporating strategies to improve safety, while Sayer and colleagues (2010) showed that Veterans whose friends and family had noticed their engaging in aggressive or dangerous driving were four times as likely to screen positive for PTSD.

3.2.2. Driving Errors

Six studies found that PTSD was related to unintentional driving errors (see PTSD/Driving-Related Findings column of Table 1). In two studies using driving simulators, Classen and colleagues (2011) noted that individuals with PTSD committed more adjustment-to-stimuli errors relative to controls, while Amick and colleagues (2013) similarly identified a trend whereby greater PTSD symptom severity was linked to higher number of total errors. Self-report work by Baker and colleagues (2014) and Clapp and colleagues (2019) showed that greater PTSD symptom severity was associated with exaggerated driving safety/caution behaviors. Both Lew and colleagues (2011) and Mairean (2020) determined PTSD was linked to more frequent lapses in attention while driving, near-misses, and other unintentional errors that may endanger other road users.

3.2.3. Driving Cognitions and Emotions

Multiple studies (k=6) explored the relationship between PTSD and thoughts or emotions experienced while driving. According to both Whipple and colleagues (2016) and Zinzow and colleagues (2013), individuals with PTSD reported higher levels of anxiety and/or hyperarousal while driving relative to healthy controls. Congruently, Lew and colleagues (2011) and Tessier and colleagues (2017) showed that individuals with PTSD endorsed greater anger and were less likely to employ coping self-instructions while driving as compared to those without PTSD. Hannold and colleagues (2013) used structured interviews to identify common triggers of feelings of anxiety (e.g., driving a small vehicle, roadside objects) and anger (e.g., unsafe passing, being followed too closely) while driving in Veterans with PTSD. More recent work by Mairean (2019) found that greater PTSD symptom severity was associated with more frequent thoughts behind the wheel related to panic (e.g., “my heart will stop beating while driving”), accident (e.g., “I will die in an accident”), and social factors (e.g., “people will think I’m a bad driver”).

3.2.4. Accidents and Legal Issues

The literature regarding associations between PTSD and motor vehicle accidents, driving-related legal issues and drinking and driving (k=5) was mixed (see PTSD/Driving-Related Findings column of Table 1). Whereas Saberi and colleagues (2013) found that those who were suspected to have PTSD reported more accidents, Alavi and colleagues (2017) did not find this to be the case after accounting for other psychiatric variables. Congruent with Alavi, Lew and colleagues (2011) identified no differences between those with and without PTSD for risk of receiving a warning or ticket or being involved in an accident, even though those with PTSD reported higher prevalence and greater persistence of driving problems overall. Finally, whereas Hoggatt and colleagues (2015) showed that those with PTSD were at increased odds of drinking and driving, Kuhn and colleagues (2010) noted that PTSD symptom severity did not predict frequency of driving following alcohol or substance use.

3.2.5. Driving Status

Two studies examined whether PTSD symptom severity was linked to active driving status, both in the context of remote TBI. After accounting for other common predictors of driving status (e.g., sensory impairment, substance use, functional impairment), Bernstein and colleagues (2021) showed that greater PTSD symptom severity was associated with lower likelihood of remaining an active driver in both those with a mTBI and those with a moderate-to-severe TBI. In univariate models, Caplan and colleagues (2017) similarly found that greater PTSD symptom severity was linked to lower likelihood of remaining an active driver in those with mTBI, however did not identify such a relationship in those with moderate-to-severe TBI.

4. Discussion

Human factors, including psychiatric disorders, are responsible for the majority of motor vehicle accidents worldwide (Petridou & Moustaki). Driving is important for most individuals’ mobility and functional independence, however this benefit must be weighed against its costs in the form of MVA-associated death and disability (Burkhardt, 1999; Ragland et al., 2005). Exposure to traumatic events as well as symptoms of and factors related to PTSD, including alterations in mood and arousal and cognitive difficulties, have been hypothesized to undermine driving-related challenges (Lew et al., 2010; Possis et al., 2014). The present review is among the first to provide a synopsis of methodologies used and findings derived from the small but emerging literature on the effects of PTSD on driving behaviors.

4.1. Methodologies

The vast majority of studies identified in this review used a cross-sectional design, whereby PTSD symptoms and driving behaviors were both assessed at a single time point. These studies identified associations between PTSD symptom severity and a wide range of driving behaviors, including hostile/aggressive driving (Clapp et al., 2014), driving errors (Amick et al., 2013) driving-related cognitions and emotions (Mairean et al., 2019), and accidents/legal issues (Alavi et al., 2017). In contrast, only three studies included any follow-up assessments and had a more narrow set of outcome variables. Two of the three studies focused on changes in driving status (Bernstein et al., 2021; Caplan et al., 2017), while the third primarily examined aggressive driving behaviors (Baker, Litwack, Clapp, Beck, & Sloan, 2014).

Notably, only Baker and colleagues (2014) examined concurrent changes in PTSD and driving behaviors. While logistically cross-sectional designs often put a lower burden on participant and research staff time and data is more readily collected, they also prohibit appreciation for longitudinal changes in predictor and outcome variables, as well as changes in the relationships between these variables over time. PTSD is a clinically heterogeneous disorder, both with respect to the specific symptoms experienced as well as its onset and course (Bonanno & Mancini, 2012). As symptoms may fluctuate from year-to-year, month-to-month or even week-to-week (Biggs et al., 2019; Bohnert et al., 2018; Solomon & Mikulincer, 2006) and temporal increases in symptoms are linked to concurrent increased difficulties completing various everyday activities (e.g., work/school tasks, community participation) (Fortenbaugh et al., 2020), it stands to reason that rises in PTSD symptom severity over time may also be associated with more pronounced disruption to facets of driving. Future studies designed to repeatedly and conjunctively assess PTSD symptom severity and driving behaviors would provide essential information on this topic, which might lay the groundwork for intervention efforts.

4.1.1. PTSD Assessment

Half of the studies employed the PCL as a measure of PTSD symptoms. The PCL is well validated and the most commonly used self-report tool in the assessment of PTSD (Blevins et al., 2015; McDonald & Calhoun, 2010). However, its diagnostic accuracy may be affected by spectrum and bias, and a structured interview (particularly the CAPS) is considered the gold standard when diagnosing PTSD (McDonald & Calhoun, 2010). Despite this, only four studies used a structured interview to determine PTSD status, including three that used the CAPS.

Additionally, while the PCL and CAPS allow for examination of PTSD symptom cluster scores, only half of the 14 identified studies using either of these measures explored the relationship between PTSD symptom clusters and driving variables. While some studies used correlation analyses to demonstrate broad associations with driving variables across clusters (Mairean et al., 2018), others used more sophisticated analytical approaches (e.g., regression models) to highlight hyperarousal (Clapp et al., 2014; Clapp et al., 2019; Kuhn et al., 2010) as being particularly strongly associated with driving-related variables. Dysphoria (Bernstein et al., 2021) and avoidance (Kuhn et al., 2010) were also selectively linked to driving variables, From a clinical standpoint, understanding the specific types of symptoms most predictive of driving difficulties may help identify those at risk for MVA-related death and disability, as well as guide intervention efforts. For instance, should higher avoidance scores be uniquely associated with reduced driving frequency (potentially due to a resistance to traveling to places or individuals that evoke memories of the traumatic event), efforts to reduce avoidance symptoms in particular may be accompanied by increased mobility within one’s community.

A total of eight studies used either the PC-PTSD as a clinician-administered PTSD screener measure or a pre-existing chart diagnosis of PTSD. While the PC-PTSD has demonstrated good diagnostic accuracy, it is by definition less exhaustive than the PCL and CAPS, and provides relatively little information about symptom severity (Prins et al., 2016). Similarly, a chart diagnosis in and of itself is not informative about the specific types and severity of symptoms experienced. Diagnoses derived from the electronic medical record have also been shown to frequently differ from those established on diagnostic interview (Holowka et al., 2014), possibly because they are often out of date.

Two studies recruited samples whose Criterion A trauma was due to motor vehicle accidents (Baker et al., 2014; Clapp et al., 2014), with both indicating that greater PTSD symptom severity was linked to higher rates of aggressive driving. Despite this, no studies evaluated how different sources of Criterion A trauma may have contrasting effects on driving behaviors. For example, it is quite possible that an individual whose Criterion A trauma stemmed from a motor vehicle accident may endorse greater interference in driving than an individual whose Criterion A trauma was due to another cause. It is hypothesized that such individuals may demonstrate greater avoidance of driving and/or be more likely to have stopped driving entirely, as well as show heightened hypervigilance or aggression toward other road users while driving, irrespective of whether they are involved in more accidents. However, in the absence of sufficient literature exploring this topic, this would be mere speculation. Future studies should use case-control designs to examine if and how such individuals differ in their driving behaviors from other drivers with PTSD whose Criterion A trauma was not due to a motor vehicle accident. In sum, more comprehensive, fine-grained assessments of the source and chronicity of trauma exposure and the types of subsequent PTSD symptoms experienced would significantly enhance understanding of the relationship between PTSD and driving behaviors.

Only two studies examined the independent effects of PTSD and TBI on driving, both of which showed that PTSD was more strongly associated with driving variables than TBI history. The driving literature as it pertains to TBI is expansive, with an array of studies and reviews covering injuries of varying temporalities (e.g., acute vs. remote) and severities (i.e., mild vs. moderate/severe) (Bernstein & Calamia, 2018; Classen et al., 2009; Liddle et al., 2014; Novack et al., 2021; Preece et al., 2010). The PTSD/driving literature is significantly smaller by comparison, with the present review representing one of the first to synthesize this body of work. Future studies exploring driving behaviors in TBI samples are advised to include PTSD as a predictor variable of interest, or at minimum to control for its effects. Moreover, given that PTSD and TBI have been shown to have a cumulative effect on aspects of daily functioning (Amick et al., 2018; Lippa et al., 2015), studies exploring whether a similar interaction exists with respect to driving behaviors are warranted (Lew et al., 2011).

4.1.2. Driving Assessment

Of the 22 studies included in the review, 20 used a subjective driving outcome measure. Collectively, these studies showed that PTSD symptom severity is associated with several aspects of driving behavior, including hostile/aggressive driving, driving errors, driving-related cognitions and emotions, and accidents/legal issues. In contrast, the small minority of studies using objective measures (Amick et al., 2013; Classen et al., 2011) employed simulator paradigms. These studies revealed associations between PTSD symptom severity and driving errors, including speeding and adjustment to stimuli. The overwhelming popularity of self-report driving measures, as opposed to objective ones, is likely in part due to their time and cost efficiency (Bernstein & Calamia, 2019). However, they have several shortcomings, including reporter biases (e.g., social desirability, selective memory), assessment of a relatively narrow set of driving-related skills/behaviors, and having only modest correlations with objective driving measures (e.g., driving simulators) and on-road performance (Freund et al., 2005; Paulhus, 1984; Schwebel et al., 2007). Furthermore, almost half of studies using subjective driving measures employed either a single item with a binary response (i.e., yes/no) or a small set of items (i.e., six or less) from larger self-report measures tapping broader constructs (e.g., risk-taking). The brevity of these measures suggests that they may not tap a sufficiently broad number of facets of driving and/or may not assess any one facet in great depth. This is a significant shortcoming given the complexity of driving as an outcome variable. Overall, by relying almost exclusively on individuals’ perceptions of their driving and using potentially insensitive measures to do so, extant literature may misestimate the true effect of PTSD on driving behaviors.

4.2. Associations Between PTSD and Driving

Although all studies were required to include an assessment of the relationship between PTSD and a driving variable in order to be included in this review, only half of the studies outlined this as an explicit goal. The paucity of literature specifically addressing this topic is surprising given that most studies recruited Veteran cohorts, a population at higher risk for both PTSD and motor vehicle accidents relative to civilians (Fulton et al., 2015; Lincoln et al., 2006). Motor vehicle accidents also increase significantly from pre to post-deployment, suggesting that sequelae secondary to deployment are uniquely predictive of driving-related difficulties during reintegration (Lincoln et al., 2006). As might be expected, studies that did not list examining the PTSD/driving relationship as an aim were in most cases the same studies that employed only brief or single-item measures of driving variables and/or relied on less comprehensive measures of PTSD, with the associated methodological shortcomings noted previously.

Half of all studies reviewed included an examination of associations between PTSD and aggressive or hostile driving behaviors, making it the driving domain most frequently assessed (see Table 2). Across these studies, individuals with PTSD and/or greater PTSD symptom severity endorsed higher rates of hostile driving behaviors. While the severity and dangerousness of the behaviors assessed varied among studies, their agreement suggests that PTSD is associated with a tendency to drive in a manner that may endanger other road users. Likely related to this point, PTSD was also consistently linked to more frequent experiencing of negative thoughts and emotions while behind the wheel, including anger, anxiety, hyperarousal, and panic. As such, it is possible that observed hostile driving behaviors represent a behavioral manifestation of emotional and/or arousal-related dysfunction (Lew et al., 2011). Considering that individuals with PTSD were also less likely to use emotional coping strategies while driving (Tessier et al., 2017) and an extensive literature suggesting that incorporation of such strategies reduces PTSD symptoms more broadly (Hassija et al., 2012; Pietrzak et al., 2011), future studies should explore whether interventions promoting the development and use of coping strategies while driving may help improve safety. For example, efforts to promote emotional regulation and progressive relaxation techniques may reduce feelings of anger and anxiety, which may in turn lead to reduced rates of hostile driving in this population. A small number of studies have also shown that PTSD treatments may be designed to particularly benefit those whose trauma or post-traumatic symptoms may be related to a motor vehicle accident. For example, Beck and colleagues (2007) developed a virtual reality exposure therapy to reduce PTSD symptoms in individuals who were involved in a motor vehicle accident, while Beck and Coffey (2005) outlined a group cognitive behavioral treatment for PTSD for such individuals (Beck et al., 2007; Beck & Coffey, 2005). Others have proposed treatments to actually improve the driving behaviors of individuals with combat exposure (Classen et al., 2014). Overall, this body of work suggests that the deleterious effects of PTSD symptoms on driving behaviors may be lessened in response to therapeutic interventions.

PTSD was also linked to more frequent unintentional driving errors, such as lapses in attention, exaggerated caution behaviors and difficulties adjusting to stimuli. Unlike the hostile behaviors summarized previously, these driving errors are usually made without conscious awareness of the mistake (Wood et al., 2013). This body of literature suggests that some PTSD symptoms may have a subtle but deleterious relationship with abilities necessary for safe driving. Thus, while individuals with PTSD are more likely than those without PTSD to engage in hostile maneuvers behind the wheel that endanger themselves and other road users, they are also at increased odds of making inadvertent mistakes behind the wheel, which may nonetheless also pose risk to others. Targeted efforts to reduce the frequency of these errors, such as through visual search retraining (Classen et al., 2017) are urgently needed.

Literature concerning the relationship between PTSD and motor vehicle accident rates/legal issues was conflicting, with some studies suggesting that PTSD was associated with higher motor vehicle accident rates and more legal involvement (e.g., tickets, drinking and driving), and others failing to identify any such relationships. These mixed results are surprising given the consistency in findings as it relates to hostile driving behaviors, negative driving thoughts/emotions, and unintentional errors. However, it is worth noting that relative to other areas of investigation, the body of work exploring motor vehicle accidents and legal repercussions of PTSD was considerably smaller (k=5), and studies varied in the specific driving outcomes assessed and extent to which psychiatric and demographic variables were controlled for. A number of reasons may explain the dearth in studies in this area, such as reporter biases and difficulties cross-validating the veracity of these claims against government records. Furthermore, it is possible that frequency of hostile driving behaviors, driving errors and negative driving thoughts/ emotions may not highly correlate with rates of accident or legal involvement, thus explaining inconsistencies in findings for the latter area of research. For example, as noted previously, some individuals may develop coping strategies that prevent PTSD-related symptoms experienced behind the wheel from influencing their driving behaviors. The comparative low base rate of accidents and legal issues may also prevent significant associations from being observed. Overall, it is premature to conclude whether and the extent to which PTSD is related to risk of motor vehicle accident and driving-related legal issues; future work should include measures assessing these critically important constructs.

4.3. Limitations

This review is subject to several limitations. Premorbid factors (e.g., risk-taking tendencies prior to trauma exposure) cannot be ruled out as potential contributors to driving behaviors (Bell et al., 2000). Given the dearth of longitudinal studies noted previously, the capacity to appreciate the temporal relations between symptom onset and driving behavior changes is presently limited. This reduces our capacity to comment on the directionality of PTSD/driving relationships. Future work should employ more complex and longitudinal models, as such findings would have important implications for clinical intervention. All studies included in the review also identified associations between PTSD symptom severity and at least one aspect of driving, a universal trend which may partially owe to publication bias.

The extant literature is also limited by generalizability concerns, as a majority of studies recruited current or former service members. Given that members of this population differ from non-military samples in several clinically meaningful ways (McGlinchey et al., 2017), there is concern that trends identified in this review may not apply to civilian samples. While the small number of civilian studies included in this review suggest that the effects of PTSD symptoms on driving may extend to non-military personnel, those studies that included both civilian and military samples (Amick et al., 2013; Classen et al., 2011) did not assess whether PTSD/driving relationships were of similar comparative strength within each of these two subgroups. Greater work would clarify the extent to which these effects may be moderated by military status. Studies also predominantly recruited from Western communities, which may limit generalizability to samples drawn from other parts of the world. For example, as countries differ significantly in the frequency with which aggressive driving is endorsed (Ozkan et al., 2011), replications of these studies in non-Western countries would increase understanding of the extent to which observed PTSD/driving relationships may be observed cross-culturally.

Finally, although this review included an assessment of the contributions of comorbid TBI to driving, other psychiatric and somatic conditions frequently accompany PTSD, such as depression, anxiety, impaired sleep, chronic pain and alcohol/substance abuse (Kobayashi et al., 2007; McGlinchey et al., 2017; Richardson et al., 2017). In conjunction with PTSD, symptoms of these conditions have demonstrated additive effects on aspects of daily functioning, such as employment status and disability (Amick et al., 2018; Lippa et al., 2015). While beyond the scope of this review, the relative dearth of reference to these comorbid conditions in the PTSD/driving literature suggests need for further exploration.

Highlights.

• Twenty-two studies examined associations between PTSD and driving behaviors

• Most studies were cross-sectional and used subjective driving measures

• PTSD symptoms linked to increased hostile driving and unintentional errors

• PTSD symptoms also associated with negative driving-related thoughts/emotions