Comparison of On-Road Driving Between Young Adults With and Without ADHD

Abstract

Objective: This study compared video recordings from routine driving of ADHD and non-ADHD young adults to identify differences in driving behaviors. Method: A matched sample of young adult drivers with and without ADHD are compared via blinded ratings of videoed g-force events recorded by DriveCam technology over 3 months of on-road driving. Results: ADHD drivers were significantly more likely to have more crashes, minor events, and g-force events. G-force events for the ADHD drivers involved significantly more risky and illegal, hyperactive/impulsive, and distracted behaviors. The g-force events of non-ADHD drivers were due to evasive, defensive driving or lapses in attention. Conclusion: Increased risk for ADHD drivers may be the result of increased risk taking, increased hyperactivity/impulsivity or distraction behavior, and increased vulnerability to factors that interfere with driving in general, whereas the consequences of faulty driving were either higher or potentially higher in those drivers with ADHD.

Keywords

ADHD driving young adults

Introduction

ADHD is associated with increased vehicular collisions, citations, and related injuries (Barkley, 2004; Barkley & Cox, 2007; D. J. Cox & Taylor-Davis, 2009; Jerome, Habinski, & Segal, 2006). Novice drivers have higher rates of accidents in their first 6 months of driving, but then the accident rate slowly declines over a number of years eventually reaching generally low adult rates around age 25 (Mayhew, Simpson, & Pak, 2003; McCartt, Shabanova, & Leaf, 2003; B. G. Simons-Morton et al., 2011a; Williams, 2003). There is some evidence suggesting that unlike patterns found in the general public, accident rates do not decline in adulthood among male ADHD drivers (Kay, Michaels, & Pakull, 2009). Although it makes sense that the various cognitive and behavioral difficulties demonstrated in ADHD would negatively affect automobile driving, this has never been directly examined in actual day-to-day driving to determine what kind of problems occur and when.

This study is the first to examine the on-road driving patterns of young adult drivers with and without ADHD that occur during g-force events while driving. G-force events are determined by in-car technology that records rapid changes in speed and direction. Elevated g-force events, one measure of risky driving, predict accident rates in novice drivers (B. G. Simons-Morton, Zhang, Jackson, & Albert, 2012). Many of the early driving errors by novice drivers are based on judgment errors (McKnight & McKnight, 2003). Adolescent drivers are also known to take greater driving risks (Williams, 2003). In addition, the social and environmental context of driving has an impact on driving safety in novice adolescent drivers. For instance, the presence of adult passengers decreases risky driving, whereas the presence of other risk-taking adolescents greatly increases risky driving (Simons-Morton et al., 2011a). On the other hand, accidents by adult drivers are overwhelmingly related to loss of attention (Dingus et al., 2006). Little is known about the exact behavioral factors that further increase the risk of young adult drivers with ADHD above those of other drivers. For instance, do drivers with ADHD make the same driving errors or have the same kinds of difficulties that young adults without ADHD have, but just more of them? Or do they make errors or have difficulties that are relatively unique to this population? Do ADHD drivers have predominately driving errors based on risk taking, judgment, or inattention, compared with age-matched, non-ADHD drivers?

Various evidences have so far suggested that attentional problems and increased risky behavior may contribute to the increased rate of accidents in young drivers with ADHD. For instance, it has been shown that adolescents with ADHD are more likely to drive under the influence of alcohol and have a higher arrest rate for such behavior (Woodward, Fergusson, & Howard, 2000). Adolescent drivers with ADHD are more likely to receive traffic citations for reckless driving, driving without a license, and driving with a revoked of suspended license (Barkley & Cox, 2007) indicating increased risk taking. Furthermore, attentional deficits may also play a large role. Through filling out daily questionnaires about their driving, Rosenbloom and Wultz (2011) demonstrated that ADHD drivers reported committing more “faults” in driving than matched non-ADHD drivers, but not “violations.” Faults consisted of driving errors, such as braking too quickly, misreading road signs, or hitting something while reversing. Violations are intentional behaviors that are illegal, such as ignoring traffic signs, illegal parking, speeding, tailgating, and not wearing a seat belt. They attribute the difference to the fact that the errors relate more to attentional problems, whereas ADHD drivers know the rules of driving and therefore are no more likely to intentionally violate driving rules. Unfortunately, self-reported instances of driving errors may be open to bias. The average age of their participant was 25.5 years, ranging from 18 to 34 years. Thus, they were not by and large novice drivers. A driving simulator based study showed that young adult drivers with ADHD (M age = 21.5 years) had significantly more deviation in lane position and increased steering rate compared with age-matched controls (M age = 22 years), but not driving speed variation (Weafer, Camarillo, Fillmore, Milich, & Marczinski, 2008). Lane position deviation and increased steering rate may indicate attentional difficulties, rather than intentional risky behaviors. This study was designed to assess on-road driving behavior through the analysis of g-force events in a sample of young adult drivers with ADHD, selected for a history of driving difficulties, compared with a matched sample of drivers without ADHD.

Materials and Methods

Participants

The study included 17 participants with clinically verified ADHD (D. J. Cox et al., 2012) and 19 non-ADHD participants matched for age, gender, race, and education level.

The inclusion criteria for ADHD participants in the study were as follows: (a) current diagnosis of ADHD as confirmed by a structured psychiatric interview and response on the Barkley questionnaire (Barkley, 2005) and the Conners’ Adult ADHD Rating Scale (CAARS; Erhardt et al., 1999); (b) not currently taking psycho-active medication; (c) having no current diagnosis of substance abuse, bipolar disorder, depression, anxiety, or psychosis as determined by the Structured Clinical Interview for the DSM-IV (SCID); (d) having a valid driver’s license; (e) routinely driving a car more than three times a week; (f) having had more than one driving mishap (collisions and/or citations) in the past 2 years; and (g) being the primary driver of a single vehicle in which a video recording system could be installed. No participants were excluded for Points 4 to 7. The inclusion criteria for non-ADHD participants in the study were as follows: (a) no current or past diagnosis of ADHD as confirmed by a structured psychiatric interview and response on the Barkley questionnaire (Barkley, 2005) and the CAARS; Erhardt et al., 1999); (b) no history of taking stimulant medication; (c) not currently taking psycho-active medication; (d) having no current diagnosis of substance abuse, bipolar disorder, depression, anxiety or psychosis as determined by the SCID; (e) having a valid driver’s license; (f) routinely driving a car more than three times a week; and (g) being the primary driver of a single vehicle in which a video recording system could be installed.

Over a period of 1.5 years, 46 individuals (25 with ADHD and 21 without ADHD) met these inclusion criteria and signed an Institutional Review Board (IRB)-approved consent form. Of the 46 participants who consented, during screening, one participant was dropped as a result of meeting the criteria for a diagnosis of bipolar disorder, one participant was dropped because of the diagnosis of active depression, another participant was dropped because he or she had lost access to his or her car, and a final participant was dropped because he or she was moving away. Four other participants declined participation due to time constraints. During video data collection, 2 ADHD participants dropped out. Thirty-six participants completed all elements of the study. See Table 1 for comparison of ADHD and non-ADHD participants.

Table 1.

Comparison of Participants.

	Age	Race (%)	Gender (%)	School (%)
ADHD (n = 17)	M = 20.71	Non-White = 23.5	F = 23.5	Nonschool = 29.4
		White = 76.5	M = 76.5	School = 70.6
Non-ADHD (n = 19)	M = 21.16	Non-White = 21.1	F = 21.1	Nonschool = 21.1
		White = 78.9	M = 78.9	School = 78.9

Procedure

Advertisements for the study were placed in local public and college newspapers and fliers were posted in public locations. Participants were offered a free, focused psychiatric and physical exam, ADHD diagnosis assessment, and up to US$300 for completing the study (with partial payment for partial completion of the study). Qualifying potential participants were informed about study procedures and were encouraged to ask questions. They were shown a picture of the in-car video monitoring system that was to be installed in their car (www.DriveCam.com, see Figure 1), and were asked to review and sign an IRB-approved consent form. Subsequently, participants were administered the SCID by a trained examiner to rule out the presence of any comorbidities, and then completed the CAARS to establish or rule out a diagnosis of ADHD. The Barkley Structured Interview for ADHD (Barkley, 2005) was also administered to further confirm an ADHD diagnosis. In addition, they completed the Cox Assessment of Risky Driving Scale (CARDS; Cox & Cox, 2009; D. J. Cox & Taylor-Davis, 2009). Participants met with a psychiatrist for medical history, physical, and neurological examinations.

Figure 1.

DriveCam in-car audio–video recording system mounted behind rearview mirror.

A DriveCam audio and video system was installed in each participant’s car behind the rearview mirror, and over the next 3-month condition driving was monitored using this system (see Figure 1). The DriveCam system consists of two video cameras and an accelerometer mounted unobtrusively behind the rearview mirror (see Figure 1). One camera faces out toward the road, capturing the driver’s view, and a second camera faces in toward the car, capturing the driver and any passengers. While the cameras were continually recording, only 10 s before and 10 s after the accelerometer detected a change in g-force of 0.6 or greater were saved. Thus, the system documents erratic driving events such as when the vehicle suddenly decelerates, accelerates, goes over a bump or swerves right or left, and so on. The DriveCam system did not indicate to the driver when the accelerometer detected an erratic maneuver and when data were being saved. All participants met monthly with a psychiatrist who recorded their body weight, heart rate, and blood pressure, and downloaded images from each participant’s DriveCam.

These 20-s videos were coded by coders blind to diagnosis using a coding system developed initially at Virginia Polytechnic Institute and State University (Dingus et al., 2006). This coding system was modified for DriveCam data at the University of Iowa (McGehee, Raby, Carney, Lee, & Reyes, 2007), and then further refined at UVa for use with drivers with ADHD (D. J. Cox et al., 2008). Data coders were instructed on the use of this coding system and practiced on pilot DriveCam data until there was 90% agreement in the coding results. The primary coder coded all videos and these data were used for data analysis. A secondary coder coded 20% of the videos. To assess interrater reliability, intraclass correlation coefficients (ICCs) were calculated between the two coders for continuous variables, and kappas were calculated for categorical variables. Variables with an ICC > .70 or kappa > .60 were considered sufficiently reliable for data analysis. To assess intrarater test–retest reliability, 45 of the original video clips (15 from controls, 15 from ADHD drivers off medication, and 15 from ADHD drivers on medication) were rerated approximately 18 months later using the same standardized rating system by the original rater who was blind to diagnosis/no-diagnosis of ADHD and medication/no-medication conditions. The mean kappa reliability for these variables was .72, which is considered a substantial agreement. The number of collisions, driver’s fault, or other’s fault were compared using a Pearsons chi-square comparing the collisions between 17 ADHD drivers and 19 non-ADHD drivers. Other in-vehicle driving data collected via DriveCam were analyzed with chi-square test and t test. Chi-square was used to compare the frequencies of driver behaviors between the ADHD drivers and normal young adult drivers groups. For two-group comparison of means—number of g-force events, g-force magnitudes, and passenger variables—t test was used.

Results

Comparison of Collisions and G-Force Events

There were eight collisions among the ADHD drivers and only one collision among the non-ADHD drivers (p = .004). Collisions were defined as a participating driver experiencing a collision with any object that resulted in physical damage to the driver’s vehicle. The collisions among the ADHD drivers consisted of running into a lead car stopped at an intersection (n = 3), backing into a stationary object (n = 2), hitting something when overcorrecting (n = 1), hitting something when turning (n = 1), and hitting a deer (n = 1). In three of these accidents, the drivers were using a cell phone (two of whom were texting), four were looking down or out a side window, appearing not to be paying attention, and in one case, it was unclear what the driver was doing. The one collision among the non-ADHD drivers consisted of the driver being stopped at a traffic light and being rear-ended by another vehicle (Table 2). The ADHD drivers were judged to be at fault in seven of the eight accidents, whereas the one accident among the non-ADHD drivers was judged to not be the driver’s fault, resulting in the ADHD drivers having a significantly greater chance of being at fault in accidents (p = .002).

Table 2.

Crashes and G-force Events.

Variable	Driving event	Non-ADHD frequency	ADHD frequency	p level
Collisions		1	8	.004^a
Driver’s fault		0	7	.002
Other’s fault		1	1
Minor events		32/649 (4.9%)	88/1,590 (5.5%)	.017
Near misses (types)	Type B: other acted	1/649 (1.6%)	10/1,590 (3.4%)
	Type A: agent acted	22/649 (3.4%)	34/1,590 (2.1%)
	Taps	9/649 (1.4%)	44/1,590 (2.8%)
Crash type		32/649	89/1,590	.007
	Lane departure	8/649 (1.2%)	4/1,590 (0.3%)
	Intersection	4/649 (0.6%)	37/1,590 (2.3%)
	Backing up	1/649 (0.2%)	2/1,590 (0.1%)
	Going straight	18/649 (2.8%)	37/1,590 (2.3%)
	Hit by someone else	1/649 (0.2%)	8/1,590 (0.5%)
G-force events		649	1,590	.000
G-force	Forward (M)	−0.1918 (SD 0.49997)	−0.1218 (SD 1.37498)	.078
	Lateral (M)	0.1983 (SD 0.61135)	0.0676 (SD 0.64457)	.000
	Braking abruptly	77 (11.9%)	202 (12.7%)	.000
	Accelerating abruptly	5 (0.8%)	7 (0.4%)
	Turning left abruptly	173 (26.7%)	572 (36%)
	Turning right abruptly	370 (57%)	758 (48%)
	Bumps	12 (1.8%)	40 (2.5%)
	Other	11 (1.7%)	11 (0.7%)

Both collisions and driver’s fault calculated using Pearson chi-square.

In addition to collisions, minor incidents were also examined, which included “near misses” and “taps” (which were defined as hitting an object but no damage incurred; Table 2). The type of minor incidents were significantly different between ADHD drivers and the non-ADHD drivers (p = .017), with the ADHD drivers having more “near misses” in which the other driver had to act to avoid them, as well as “taps” compared with the non-ADHD drivers.

Crash type (Table 2) refers to the nature of the driving when minor events and collisions occurred. The overall pattern of these was significantly different between the ADHD and the non-ADHD drivers (p = .007). The ADHD drivers were four times more likely to be driving through an intersection when the events occurred, whereas the non-ADHD drivers were four times more likely to make a lane departure than the ADHD drivers, suggesting an evasive maneuver.

There were 649 g-force events in the non-ADHD drivers and 1,590 g-force events in the ADHD drivers (p < .000; Table 2). The mean strength of the Lateral g-force events was significantly stronger among non-ADHD drivers than among the ADHD drivers (p = .000), whereas the mean strength of the Forward g-force neared significance (p = .078) with the non-ADHD drivers again having stronger g-force (Table 2). The (−) sign in the Forward g-force refers to the direction of the force, thus indicating an abrupt stop. The pattern of event types differed significantly between the ADHD drivers and the non-ADHD drivers (p = .000), but the percentage of each subcategory of events did not differ between the two sets of drivers, thus it was the overall pattern of events that was significantly different.

Internal and External Factors in G-Force Episodes

The g-force events were further categorized as to contextual (in car) and environmental factors (outside car; Table 3). Context characteristics include a passenger being present in the car and interaction with the passenger, while environmental characteristics include darkness and weather. ADHD drivers were significantly more likely to have another young adult in the car (p = .000), as well as being significantly more likely to be interacting with a passenger (p = .000) during the g-force event. ADHD driver g-force events were significantly more likely to occur in poor weather (p = .14) or in darkness (p = .000).

Table 3.

Internal and External Factors in G-Force Episodes.

External environment
Weather	Poor weather: visible fog, rain, snow, sleet, or hail	25/649 (3.9%)	113/1,589 (7.1%)	.014
Light	Darkness (streetlights on, headlights being used my most cars)	221/649 (34.1%)	690/1,589 (43.4%)	.000
Internal context
Passenger 1	Another young adult; more than one passenger either all young adult or all children	88/649 (13.6%)	347/1,588 (21.9%)	.000
Passenger 2	Interaction with passenger (conflict, conversation, or physical action by passenger)	89/649 (13.7%)	347/1,588 (21.8%)	.000

Comparison of the Drivers’ Behavior in the G-Force Events

G-force events were divided into subtypes based on analysis of drivers’ behaviors—violations and errors, in part based on Rosenbloom and Wultz (2011), and reactions to events. Violations are defined as “deliberate departures from behaviors believed to represent safe driving practices” (Rosenbloom & Wultz, 2011, p. 128), whereas errors are defined as “failures of observation that may be hazardous to others” (Rosenbloom & Wultz, 2011, p. 128). We further divided errors into those that appeared to be related to impulsivity or hyperactivity and those related to inattentiveness. Violations included speeding, reckless driving, use of drugs or alcohol, and the driver or one or more passengers not wearing seat belts. Errors due to impulsivity or hyperactivity included not having hands on the wheel, reaching for a moving object in the car, or being hyperactive while driving. Errors due to inattention include daydreaming, fatigue, eyes off the road, looking at objects outside the car, adjusting devices in the car, grooming, reading, using a cell phone, dining, and reaching for lighting, smoking, or extinguishing a cigarette while driving. The driver’s reaction to the event included either enjoyment or appearing scared or distressed, or the driver talking to his or her self.

As to potentially illegal driving behavior, ADHD drivers were significantly more likely (p = .000) to be unbuckled, as were their passengers, both front and back seat (p = .000, for both; Table 4). ADHD drivers were also significantly more likely to be judged to be driving reckless (p = .019) and possibly speeding (p = .000).

Table 4.

Drivers’ Behavior.

Variable	Driver behavior	Non-ADHD frequency	ADHD frequency	p level
Violations	No seat belt—driver	82/649 (12.6%)	529/1,589 (33.3%)	.000
	No seat belt—front passenger	28/649 (4.3%)	149/1,584 (9.4%)	.000
	No seat belt—back passenger	16/649 (2.5%)	80/1,584 (5.0)	.000
	Speeding	15/649 (2.3%)	289/1,589 (18.2%)	.000
	Reckless driving	4/649 (0.6%)	38/1,587 (2.3%)	.019
Impulsive/hyperactive errors	Both hands on wheel	155/649 (23.9%)	672/1,571 (42.8%)	.000
	One hand on wheel	457/649 (70.4%)	834/1,571 (53.1%)
	Two hands off wheel	4/649 (0.6%)	32/1,589 (2%)
	Cannot tell	33/649 (5.1%)	60/1,571 (3.8%)
	Moving object in car	16/649 (2.5%)	82/1,584 (5.2%)	.014
	Hyperactive	13/649 (2%)	142/1,584 (9.3%)	.000
Inattentive/distracted errors	Inattentive/daydreaming	297/649 (45.8%)	548/1,584 (34.7%)	.000
	Eyes off road > 3 s	8/649 (1.2%)	16/1,584 (1.0%)	.068
	Adjusting car devices	10/649 (1.5%)	55/1,585 (3.5%)	.014
	Smoking	1/649 (0.2%)	41/1,584 (2.6%)	.000
	Using cell phone	27/649 (4.2%)	62/1,584 (3.9%)	ns
	No music	99/649 (15.2%)	370/1,589 (23.3%)	.000
	Music on but not singing	502/649 (77.3%)	1,109/1,589 (69.8%)
	Singing along with music	48/649 (7.4%)	110/1,589 (6.9%)
Driver’s reaction	Looks at camera	54/649 (8.3%)	271/1,583 (17%)	.000
	Enjoyment	6/649 (0.9%)	39/1,584 (2.5%)	.002
Talking to themselves	Not talking to themselves	617/649 (95%)	1,541/1,587 (97%)	.048
	Talking to themselves	32/649 (5%)	46/1,587 (3%)	.048
Talking to self	Before and driving related	14/649	10/1,584	.093
Talking to self	Before and nondriving related	5/649	11/1,584	.093

Note: ns = not significant.

Behaviors that were considered to represent impulsive and hyperactive behavior were by and large much more present with ADHD drivers: having both hands off the steering wheel (.000); reaching for some sort of moving object in the car, such as the driver reaching for a lose object in the car that fell or distracted by an animal or insect in the car (p = .014); or being hyperactive, such as dancing or moving to music while driving or being fidgety and squirming while driving (p = .000). On the contrary, examples of inattentive behavior (looking at the road but daydreaming, looking out driver’s window, looking out passenger’s window, looking at rearview mirror, or looking inside the car) were more common in the non-ADHD drivers at the time of the g-force incident (p = .000; Table 4). This pattern was also seen when time of eyes off the road was measured, such that there was a nonsignificant trend for the non-ADHD drivers to have their eyes off the road for 3 s or longer at the time of the incident (p = .068). On the other hand, ADHD drivers were more likely to be distracted by certain activities at the time of the g-force incident than were non-ADHD drivers: adjusting a device, such as the temperature or entertainment, while driving (p = .014) and smoking (p = .000). The pattern of music use was also significantly different with ADHD drivers being less likely to have music on and if it was on, less likely to be singing along (p = .000). Yet, there was no significant difference between non-ADHD and ADHD drivers in potential distraction caused by use of cell phone, reading, dining, or engaging in personal grooming while driving (Table 4).

ADHD drivers during the g-force incident were significantly more likely to look at the camera after the episode (p = .000) (Table 4). ADHD drivers were also more likely to appear to be expressing happiness, joy, or excitement at the time of the g-force event (p = .002). On the other hand, the non-ADHD drivers were more likely to be talking to themselves during the g-force incident (p = .048; Table 4). Incidents of the drivers’ talking to themselves were further divided as to whether the talking occurred before the event or afterwards and whether it was related to the incident. There was a nonsignificant trend for the non-ADHD drivers to be talking to themselves before the incident and the self-talk was more related to the impending incident (p = .093). Content examination suggested that the non-ADHD drivers were alerting themselves to a potential event.

Discussion

A sample of young adult drivers with ADHD who were specially selected because of a history of driving difficulties were compared with a matched sample of young adult drivers without ADHD over a 3-month period via DriveCam technology, which captures g-force events. The young adult ADHD drivers had significantly more at fault accidents and g-force events. Furthermore, the findings from this study evidence that there are qualitative differences between the driving events of those with ADHD and those without. The driving episodes captured by the DriveCam in the non-ADHD sample appear to be more related to defensive maneuvers or lapses of attention, echoing the findings of the 100-Car Naturalistic Driving Study (Dingus et al., 2006).

The non-ADHD drivers had a different type of g-force events compared with the ADHD drivers. This is supported by the fact that the overall pattern of event types (braking abruptly, accelerating abruptly, turning right or left abruptly, and collisions) was significantly different. In addition, the mean Lateral g-force in the events was significantly greater for non-ADHD drivers. The pattern of minor events was significantly different between the non-ADHD and the ADHD drivers, with more of the non-ADHD near misses being due to defensively avoiding another driver, whereas the ADHD drivers were more often the cause of the other driver having to evade collision. The non-ADHD drivers swerved defensively, possibly explaining the higher Lateral g-force. The non-ADHD near misses were more often the result of lane departure, which may again be attributed to the non-ADHD driver having to defensively swerve to avoid another driver. The trend for non-ADHD drivers to have more self-talk may be part of this behavior.

In contrast, the near misses of the ADHD drivers were significantly more likely to occur in intersections where divided attention is more critical. The driving context for g-force events with ADHD drivers was significantly more likely to include another young adult in the car and there was significantly more interaction with that passenger during the g-force event. Although it has been demonstrated that high risk driving and crash/near crash rates are significantly higher for teenage novice drivers with other teenage males or risk-taking friends in the car, demonstrating that risky driving is socially influenced (B. Simons-Morton, Lerner, & Singer, 2005; B. G. Simons-Morton et al., 2011b), we do not know whether this occurs more frequently with ADHD drivers or is just more likely to contribute to driving events. Nonetheless, interaction with passengers may have contributed to decreased safe driving, especially in circumstances where increased attention is required, such as in intersections. The fact that the g-force events of ADHD drivers were significantly more likely to occur in bad weather or darkness is difficult to interpret because we do not have a frequency count of the actual driving conditions for the two samples. It could be that ADHD drivers’ ability to drive safely may be more affected by adverse conditions than non-ADHD drivers, or that ADHD drivers are more likely to drive in such situations of increased risk. It may also be that ADHD drivers are less likely to alter their driving behaviors to take into consideration the increased risk entailed in adverse driving circumstances. Nonetheless, adverse driving conditions—darkness and rain—were more often a factor in g-force events for ADHD drivers. It may be concluded then that adverse driving conditions, such as darkness, bad weather, or having young passengers in the car, have a greater influence on drivers with ADHD than on those without ADHD.

During the g-force events, ADHD drivers were significantly more likely to be unbuckled, raising a number of safety concerns. As ADHD drivers are more likely to have accidents, their being unbuckled makes the potential tragic consequences of such accidents higher. Because the nonwearing of seat belts probably did not contribute to the occurrence of the g-force event, it is likely that drivers with ADHD are less likely to wear their seat belts, but it is unclear why those with ADHD are less likely to wear seat belts. There are several possible explanations. This could be deliberate risk-taking behavior. People with ADHD often take more risks in general, which may account for the increased frequency of being unbuckled. On the other hand, it may be a failure to have internalized the behavior of buckling one’s seat belt. There may be other psychological factors at play. Wearing seat belts is better predicted by a sense of self-efficacy rather than risk perception (Schwarzer et al., 2007). Adults with ADHD have been shown to have, in general, a lower sense of self-efficacy (Edel et al., 2009; Newark & Stieglitz, 2010; Norwalk, Norvilitis, & MacLean, 2009). Unfortunately we did not measure this variable. We can only speculate as to why the passengers of ADHD drivers are also less likely to be wearing seat belts at the time of g-force events, but it raises the chances for a potentially disastrous outcome even higher. ADHD drivers were also judged to be more likely speeding at the time of g-force events. We could not determine whether they were actually above the speed limit, but the fact that they were driving faster than the surrounding cars is suggestive of increased risk behavior at the time of the g-force event. Therefore, besides being more susceptible to the impact of adverse driving conditions, young adult drivers with ADHD appear to have more risky driving behavior and to increase the potential negative outcome of such risky behavior by less often wearing seat belts, which was also present in their passengers.

Impulsive/hyperactive and some inattentive behaviors were judged to be present more often in g-force events involving ADHD drivers possibly contributing to the increased number of g-force events. ADHD drivers were significantly more likely to not have both hands on the steering wheel, to be hyperactive and fidgety when driving, and to be reaching for loose objects in the car. The ADHD drivers were significantly more likely to be distracted by other activities, such as adjusting car devices and smoking, but not using a cell phone, reading, grooming, or eating and drinking. The fact that three of the accidents of the ADHD drivers occurred while using the cell phone but that ADHD drivers were not more likely than non-ADHD drivers to be using the cell phone during g-force events is interesting. This may suggest that cell phone use may contribute to driving events for both ADHD and non-ADHD drivers, but the negative consequences may be greater for ADHD drivers. This possibility coordinates with the other conclusions above, namely, that anything that interferes with safe driving will have a greater impact on those with ADHD and that there may be greater negative consequences of that interference in those with ADHD.

However, there are several pieces of data that do not fit with expectations, for instance, that non-ADHD drivers were more likely judged to be inattentive and a trend toward their being more likely to have their eyes off the road for 3 s or more. This may be understood by considering that a significantly greater proportion of g-force events in non-ADHD drivers were attributable to nonattention. The main cause of accidents in general as demonstrated in the 100-Car Naturalistic Driving Study was lack of driver attention (Dingus et al., 2006). Inattention may play a relatively larger role in the occurrence of non-ADHD g-force events, because they are less likely to be at risk due to the absence of other factors that appear to contribute to increased g-force events in ADHD drivers, such as increased risk taking and increased impulsive/hyperactive behaviors. The role of listening to or singing along with music is unclear. Although ADHD drivers were less likely to have music playing, they were also less likely to be singing along when music was present. It therefore seems to have played a larger role for non-ADHD drivers.

Interestingly, ADHD drivers and non-ADHD drivers have significantly different emotional reactions to the g-force events. ADHD drivers were more likely to look at the camera and more likely to react with happiness, joy, or excitement as opposed to fear and distress, suggesting that the ADHD drivers are less likely to take the situation seriously and maybe less likely to learn from their mistake. Non-ADHD drivers were more likely to be talking to themselves during the g-force event, and further analysis gives a trend toward more self-talk, prior to the event, about the event for the non-ADHD drivers, suggesting more alerting self-talk occurring.

This is the first naturalistic study to report direct observations of the difference between young adult drivers with ADHD and those without ADHD. Our observations support the idea that ADHD drivers not only have more accidents and more g-force events than non-ADHD drivers but there are also qualitative differences between the g-force events of ADHD drivers and non-ADHD drivers. The nature of g-force events are quite different for ADHD drivers than non-ADHD drivers, with non-ADHD drivers being more likely to have g-force events in either the situation of lapsed attention to the road or as part of defensive driving, whereas ADHD drivers have more events triggered by adverse driving conditions, increased risky driving behavior, increased impulsive and hyperactive behavior, and distraction by factors inside the car. Furthermore, the potential negative consequences of accidents by ADHD drivers are much higher, as they and their young adult passengers were less often wearing seat belts.

Not all young drivers with ADHD have significant difficulty with driving. Because we specifically recruited young adults with ADHD who had a minimum record of driving difficulty to assure a contrast with young adult drivers without ADHD, these results are not generalizable to all drivers with ADHD. The fact that the DriveCam is only activated in g-force events limits the conclusions that can be derived, because we are not able to assess whether these risk-increasing behaviors observed in ADHD drivers at the time of g-force events actually occur more frequently in ADHD drivers. However, this method is superior to recall and self-reports. Thus, whether the ADHD drivers have more behaviors that contribute to adverse events or merely that these events are more likely to result in g-force events, but occur no more frequently than seen in non-ADHD drivers, cannot be answered at this time. The next step is to duplicate this study, while recording more details of the actual driving behaviors of ADHD versus non-ADHD young adult drivers.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Shire funds paid for Quyen Nichols’s time and subject payment.

Author Biographies

Richard L. Merkel, Jr., MD, PhD, DFAPA, is an Associate Professor of Psychiatry and Neurobehavioral Sciences at the University of Virginia. He attended medical school at the University of Virginia and completed his residency, a RWJ Clinical Scholars Fellowship, and his PhD in Anthropology at the University of Pennsylvania.

J. Quyen Nichols, BA, is pursuing a doctoral degree in Clinical Psychology at the University of Vermont.

Jonathan C. Fellers, MD, studied medicine at Cornell prior to completing a psychiatry residency at UVA. He is now an addiction psychiatry fellow at OHSU with research interests in stimulant misuse and abuse.

Priscilla Hidalgo, MD, is a Board-certified psychiatrist and a Geriatric Psychiatry Fellow at the University of Virginia. After obtaining her medical degree in Dominican Republic (where she is originally from), she completed her Adult Psychiatry Residency at the University of Virginia. In addition to Geriatrics, her areas of clinical interest include Personality Disorders, Psychodynamic Psychotherapy and Transcultural Psychiatry.

Lady A. Martinez, MD, is currently one of the Child & Adolescent Psychiatry Fellows at the University of Connecticut. Lady graduated “Summa Cum Laude” from Medical School in her country of origin (Dominican Republic). She moved to the United States in the summer of 2008 to complete her Adult Psychiatry Residency at the University of Virginia. Her areas of clinical interest include: externalizing disorders, Complex PTSD, and Transcultural Psychiatry. She recently earned her Board Certification in General Psychiatry.

Ivan Putziger, MD, is Chief Resident at the Department of Psychiatry and Neurobehavioral Sciences at the University of Virginia. Born in Rio de Janeiro, Brazil, and raised in Paris and New York City, he studied Neuroscience at Johns Hopkins University, and received his medical degree from Emory University.

Roger C. Burket, MD, DFAPA, DFAACAP, is a child psychiatrist and Chief, Developmental Disorder’s Section, UVA Department of Psychiatry and Neurobehavioral Sciences. He is a graduate of The George Washington University Medical School, and completed his Psychiatry and Child and Adolescent Psychiatry Training at the University of Florida. His research and clinical focus has been in the areas of eating disorders, delinquency, and ADHD.

Daniel J. Cox, PhD, ABPP, is professor of Psychiatry and Neurobehavioral Sciences, and Director of the Virginia Diving Safety Laboratory. He has extensive research experience in driving safety of various medical conditions, including Attention Deficit Hyperactivity Disorder, Diabetes, Traumatic Brain Injury.

References

Barkley

R. A.

(2004). Driving impairments in teens and adults with attention-deficit/hyperactivity disorder. Psychiatric Clinics of North America, 27, 233-260.

Barkley

R. A.

(2005). Attention-deficit hyperactivity disorder: A handbook for diagnosis and treatment (3rd ed.). New York, NY: Guilford.

Barkley

R. A.

Cox

(2007). A review of driving risks and impairments associated with attention-deficit/hyperactivity disorder and the effects of stimulant medication on driving performance. Journal of Safety Research, 38, 113-128.

Cox

B. S.

Cox

D. J.

(2009, May). Risk taking while driving: A scale to measure risk taking and its relationship to driving mishaps. Poster Session presented at the 162nd Annual Meeting of the American Psychiatric Association. Boston, MA.

Cox

D. J.

Davis

Mikami

A. Y.

Singh

Merkel

R. L.

Burket

(2012). Long acting methylphenidate reduces collision rates of young adult ADHD drivers. Journal of Clinical Psychopharmacology, 32, 225-230.

Cox

D. J.

Mikami

Cox

B. S.

Coleman

M. T.

Mahmood

Sood

Merkel

R. L.

(2008). Effect of long-acting OROS methylphenidate on routine driving in young adults with attention-deficit/hyperactivity disorder. Archives of Pediatrics & Adolescent Medicine, 162, 793-794.

Cox

D. J.

Taylor-Davis

(2009). Attention deficit/hyperactivity disorder (ADHD) and driving safety. In Verster

J. E.

Pandi-Perumal

S. R.

Ramaekers

J. G.

de Glier

J. J.

(Eds.). Drugs, driving and traffic safety (pp. 315-330). Basel, Switzerland: Birkhäuser Verlag.

Dingus

T. A.

Klauer

S. G.

Neale

V. L.

Petersen

Lee

S. E.

Sudweeks

Knipling

R. R.

(2006). The 100-Car Naturalistic Driving Study, Phase II—Results of the 100-Car Field Experiment DOT HS 810 593. National Highway Traffic Safety Administration, USDOT. Retrieved from http://www.nhtsa.gov/DOT/NHTSA/NRD/Multimedia/PDFs/Crash%20Avoidance/2002/100CarPhase1Report.pdf

Edel

M. A.

Pfutze

E. M.

Lieder

Assion

H. J.

Ribbert

Juckel

Brune

(2009). Self concept, action control and ADHD symptoms under methylphenidate treatment in adults with ADHD. Pharmacopsychiatry, 42, 109-113.

10.

Erhardt

Epstein

J. N.

Conners

C. K.

Parker

J. D. A.

Sitarenios

(1999). Self-ratings of ADHD symptoms in adults II: Reliability, validity, and diagnostic sensitivity. Journal of Attention Disorders, 3, 153-158.

11.

Jerome

Habinski

Segal

(2006). Attention-deficit/hyperactivity disorder (ADHD) and driving risk: A review of the literature and a methodological critique. Current Psychiatry Reports, 8, 416-426.

12.

Kay

G. G.

Michaels

M. A.

Pakull

(2009). Simulated driving changes in young adults with ADHD receiving mixed amphetamine salts extended release and atomoxetine. Journal of Attention Disorders, 12, 316-329.

13.

Mayhew

D. R.

Simpson

H. M.

Pak

(2003). Changes in collision rates among novice drivers during the first months of driving. Accident Analysis & Prevention, 35, 683-691.

14.

McCartt

A. T.

Shabanova

V. I.

Leaf

W. A.

(2003). Driving experience, crashes, and traffic citations of teenage beginning drivers. Accident Analysis & Prevention, 35, 311-320.

15.

McGehee

Raby

Carney

Lee

J. D.

Reyes

M. L.

(2007). Extending parental mentoring using an event-triggered video intervention in rural teen drivers. Journal of Safety Research, 38, 215-227.

16.

McKnight

A. J.

McKnight

A. S.

(2003).Young novice drivers: Careless or clueless? Accident Analysis & Prevention, 5, 921-925.

17.

Newark

P. E.

Stieglitz

R. D.

(2010). Therapy-relevant factors in adult ADHD from a cognitive behavioral perspective. Attention Deficit and Hyperactivity Disorder, 2, 59-72.

18.

Norwalk

Norvilitis

L. M.

MacLean

M. G.

(2009). ADHD symptomatology and its relationship to factors associated with college adjustment. Journal of Attention Disorders, 13, 251-258.

19.

Rosenbloom

Wultz

(2011). Thirty-day self-reported risky driving behaviors of ADHD and non-ADHD drivers. Accident Analysis, 43, 128-133.

20.

Schwarzer

Schüz

Ziegelmann

J. P.

Lippke

Luszczynska

Scholz

(2007). Adoption and maintenance of four health behaviors: Theory-guided longitudinal studies on dental flossing, seat belt use, dietary behavior, and physical activity. Annals of Behavioral Medicine, 33, 156-166.

21.

Simons-Morton

B. G.

Ouimet

M. C.

Zhang

Klauer

S. E.

Lee

S. E.

Wang

Dingus

T. A.

(2011a). Crash and risky driving involvement among novice adolescent drivers and their parents. American Journal of Public Health, 101, 2362-2367. doi:10.2105/AJPH.2011.300248

22.

Simons-Morton

B. G.

Ouimet

M. C.

Zhang

Klauer

S. E.

Lee

S. E.

Wang

Dingus

T. A.

(2011b). The effect of passengers and risk-taking friends on risky driving and crashes/near crashes among novice teenagers. Journal of Adolescent Health, 49, 587-593.

23.

Simons-Morton

B. G.

Zhang

Jackson

J. C.

Albert

(2012). Do elevated gravitation force events while driving predict crashes and near crashes. American Journal of Epidemiology, 175, 1075-1079.

24.

Simons-Morton

Lerner

Singer

(2005). The observed effects of teenage passengers on the risky driving behavior of teenage drivers. Accident Analysis & Prevention, 37, 973-982.

25.

Weafer

Camarillo

Fillmore

M. T.

Milich

Marczinski

C. A.

(2008). Simulated driving performance of adults with ADHD: Comparisons with alcohol intoxication. Experimental and Clinical Psychopharmacology, 16, 251-263.

26.

Williams

A. F.

(2003). Teenage drivers: Patterns of risk. Journal of Safety Research, 34, 5-15.

27.

Woodward

L. J.

Fergusson

D. M.

Howard

L. J.

(2000). Driving outcomes of young people with attentional difficulties in adolescence. Journal of the American Academy of Child & Adolescent Psychiatry, 39, 627-634.