Abstract
Purpose:
Adolescents with ADHD have 30–40% higher crash rates. However, we still do not understand which factors underlie heightened crash risk and if crash circumstances differ for drivers with ADHD. We compared prevalences of crash responsibility, driver actions, and crash types among adolescent and young adult drivers with and without ADHD who crashed within 48 months of licensure.
Methods:
In this exploratory retrospective cohort study, we identified patients of Children’s Hospital of Philadelphia’s (CHOP) New Jersey primary care locations who were born 1987–2000, NJ residents, had their last CHOP visit ≥ age 12, and acquired a driver’s license. We linked CHOP electronic health records to NJ’s licensing and crash databases. ADHD diagnosis was based on 13 ICD-9-CM/ICD-10-CM codes. Prevalence ratios were estimated using generalized estimating equation log-binomial regression.
Results:
We identified 934 drivers with ADHD in 1,308 crashes and 5,158 drivers without ADHD in 6,676 crashes. Within 48 months post-licensure, drivers with ADHD were more likely to be at fault for their crash (PR: 1.09 [1.05−1.14]) and noted as inattentive (1.15 [1.07−1.23]). With the exception that drivers with ADHD were less likely to crash while making a left/U-turn, we did not find substantial differences in crash types by diagnosis. Analyses also suggest females with ADHD may have a higher risk of colliding with a non-motor vehicle and crashing due to unsafe speed than females without ADHD.
Conclusions:
Results suggest crash circumstances do not widely differ for drivers with and without ADHD but highlight several factors that may be particularly challenging for young drivers with ADHD.
Keywords: automobile driving, adolescent health, attention deficit and disruptive behavior disorders, developmental disabilities, neurodevelopmental disorders, traffic accidents
Attention-deficit/hyperactivity disorder (ADHD) is one of the most common childhood disorders, with more than 1 in 10 US adolescents currently diagnosed and particularly high prevalence among males [1,2]. Becoming licensed to drive is an important rite of passage for most adolescents, dramatically increasing their mobility by providing them with opportunities to travel independently to places of employment, school, and activities. Although at rates lower than adolescents without ADHD, most adolescents with ADHD become licensed to drive by their 18th birthday [3]. However, the defining symptoms of ADHD (inattention, hyperactivity, and impulsivity) and impairments in executive functioning skills may lead to increased engagement in unsafe driving behaviors and higher rates of adverse driving outcomes [4,5]. Adolescents with ADHD frequently have impairment in response inhibition (e.g., withholding a response that is inappropriate in a specific context), interference control (e.g., ability to maintain performance despite distraction), and working memory (e.g., forethought in route planning), all of which are skills critical for driving [6]. Thus, it is important to comprehensively understand the extent to which drivers with ADHD are at higher crash risk, and subsequently why their crash risk is higher, with the ultimate goal of promoting safe driving practices and improving safe mobility among drivers with ADHD [7].
Population-based studies established that adolescents and young adults with ADHD have 30–40% higher rates of crashes and crash-related hospitalizations than those without ADHD [3,8]. A critical next step is to gain an understanding of the specific factors and mechanisms that underlie these elevated rates. Simulator studies of young adult drivers (e.g., higher maximum speed, abrupt braking) and naturalistic or on-road driving studies (e.g., elevated gravitational-force events) have shown associations between ADHD and poor driving outcomes in [9,10]. Additionally, a naturalistic driving study that included 10 adolescent drivers with ADHD found that these teens had higher rates of driving errors, including improper turns and following too closely, when compared with non-ADHD peers [11]. While these studies identify some specific driver-related proximate factors (i.e., factors occurring immediately before a crash, such as inattention) that might underlie or contribute to the heightened crash risk for adolescents with ADHD, how the prevalence of these factors differs among adolescent drivers with and without ADHD in real-world crash events remains unclear. Identifying differences in driving behaviors, errors, and circumstances of real-world crashes involving drivers with and without ADHD will directly inform the need for and development of tailored medical, behavioral, technological, policy, or educational interventions for drivers with ADHD.
Thus, as an initial step, we conducted a large, exploratory retrospective cohort study of primary care patients of the Children’s Hospital of Philadelphia (CHOP) to compare the crash circumstances (i.e., manner in which crash occurs) for crash-involved adolescent and young adult drivers (hereafter referred to as “adolescent drivers”) with and without ADHD over the first 48 months of licensure. In prior analyses, we found drivers with ADHD were more likely to experience adverse driving outcomes indicative of risky driving behavior—including alcohol-related crashes and moving violations for alcohol/substance use, not wearing a seat belt, and speeding [3,12]. In this paper, we focused on the 6,092 adolescent drivers in this cohort (934 with ADHD and 5,158 without ADHD) who were involved in a crash during their first 48 months of licensure. Specifically, we compared the prevalence of key crash circumstances, including (1) specific actions the driver committed that contributed to the crash (e.g., inattention, unsafe speed), (2) crash responsibility (i.e., fault), and (3) the specific type of crash (e.g., index driver in rear-end crash).
Methods
Data Source
Data for this study come from the New Jersey Safety and Health Outcomes (NJ-SHO) data warehouse. The NJ-SHO is a unique data source that contains linked data from numerous administrative data sources in NJ, including CHOP electronic health records (EHRs) for all patients who reside in NJ, the full licensing history of every NJ driver, and detailed information on each police-reported crash (i.e., a crash resulting in an injury or > $500 property damage) from January 2004 through December 2017. A detailed description of the individual data sources, process we undertook to probabilistically link these data sources, and results of our formal validation efforts—which indicated that all linkages were conducted with high quality—are available in a prior study [13].
Study Sample
Individuals for this retrospective cohort study were patients of the six NJ primary care practices of the CHOP pediatric network. CHOP is located in southeastern Pennsylvania and southern NJ, supports a socioeconomically, racially, and ethnically diverse population, and accepts most insurance plans, including public insurance (e.g., Medicaid). It has over a million annual visits utilizes a single, unified EHR system for all aspects of care.
The methods we used to select the sample for this cohort study are described in two previous studies [3,12]. Briefly, we queried EHR data in the NJ-SHO warehouse and identified 25,177 individuals who: (1) were born 1987–2000, (2) were patients at one of six NJ-based primary care practices, and (3) (to establish NJ residency) had a CHOP network visit (to any location) as a NJ resident within 4 years of becoming eligible for their learner’s permit at 16 years old and maintained a NJ address through their last network visit. We excluded 1,651 individuals diagnosed with an intellectual disability (n=106), only had one primary care visit (n=914), or had their last primary care visit before age 12 (n=631). Finally, we limited our population to individuals who had two or more primary care visits to minimize ADHD misclassification, and those who had their last CHOP primary care visit at age 12 or older to ensure individuals were seen by a primary care provider at an old enough age to confirm ADHD status (DSM-5 criteria calls for ADHD symptom onset by age 12). The underlying study cohort included 23,526 primary care patients (Figure 1).
Figure 1.
Flowchart depicting selection of final study cohort. Grey boxes show individuals who were excluded from the study.
ADHD Classification
Individuals were classified as having ADHD if their EHR indicated an ICD-9-CM code beginning with ‘314’ or ICD-10-CM code of ‘F90’ either at any CHOP visit or on their list of known chronic conditions. We previously validated this algorithm using manual review of EHRs; estimated sensitivity was 0.96 and specificity was 0.98 [14]. Of the 23,526 subjects in the underlying cohort, 3,368 (14.3%) were determined to have ADHD and 20,158 (85.7%) did not have ADHD (Figure 1).
Study Outcomes
This study was restricted to the 934 individuals with ADHD and 5,158 without ADHD who obtained a driver’s license, held that license for at least one full month during the study period, and were involved in at least one crash in the first 48 months of licensure (Figure 1). Data on police-reported crashes were ascertained from the NJ Police Crash Investigation Report [15]. NJ investigating officers are instructed to document on crash reports the most prominent proximate factors (at least one per crash and up to two per driver) contributing to a crash, regardless of whether a citation is issued. Possible crash-contributing factors include driver actions, vehicle factors, and road/environmental factors. For each crash, we determined whether each crash-involved driver was noted to have committed one or more actions that contributed to the crash (crash-contributing driver actions). There are 14 specific crash-contributing driver actions listed on the NJ crash report, including inattention, unsafe speed, failure to yield the right of way to vehicle/pedestrian, following too closely, and backing unsafely; our analyses focused on the most frequent driver actions (noted in >4% of study crashes). Additionally, based on our and others’ prior work on methods to determine crash responsibility [16,17], we determined a driver to be at-fault (i.e., responsible) for their crash if they were noted to have committed one or more crash-contributing driver actions. Any number of drivers—including none—could be at fault for each crash. Finally, to provide a comprehensive picture of the movement of each driver’s vehicle just prior to and at the time of the crash (i.e., crash type), we combined two variables from the crash report: (1) initial configuration of the crash, and (2) impact location on that vehicle. The initial crash configuration indicated the manner in which the vehicle first crashed into another vehicle or fixed object, such as same direction (rear-end) with another motor vehicle or collision with something other than a motor vehicle (e.g., fixed object, animal). The initial impact location describes the area of the vehicle in which the initial impact occurred; this variable was used to differentiate the index [striking] vehicle from the struck vehicle in rear-end crashes.
Other Variables
Demographic variables were derived from the EHR and included sex, race/ethnicity, insurance payor at last CHOP visit, diagnosis of anxiety disorder (ICD-9-CM beginning with ‘300’ or ICD-10-CM beginning with ‘F40,’ ‘F41,’ ‘F42,’ ‘F44,’ ‘F45,’ or ‘F48’), and diagnosis of disruptive behavior disorder (DBD) (ICD-9-CM beginning with ‘312’ or 313.81 or ICD-10-CM beginning with ‘F91’). Age at licensure was derived from licensing records using driver’s date of birth and date of licensure. Using crash reports as the primary source and licensing data as the secondary source, we geocoded the residential address of each driver to their residential census tract (ArcGIS 10.5.1, Esri, Redlands, CA). Each NJ census tracts’ median household income and population estimates were obtained from the 2013–2017 American Community Survey 5-year estimates and geographic area (square miles) from the 2010 Census Gazetteer Files, which we then used to calculate population density (population per square miles) [18,19]. Census tracts were categorized into quintiles of median household income and population density.
Statistical Analysis
We compared bivariate distributions of relevant demographic and clinical characteristics among drivers with and without ADHD using chi-square tests for categorical variables and Wilcoxon rank-sum tests for continuous variables. Using data on crashes occurring within 24 and 48 months of licensure, we compared for drivers with and without ADHD the proportion of all crashes in which the adolescent driver was at fault and the proportion in which the teen driver committed specific crash-contributing driver actions. Further, we compared the prevalence of specific crash types among at-fault crashes for drivers with and without ADHD. We then used generalized estimating equation log-binomial regression models to estimate adjusted prevalence ratios (aPR) and 95% confidence intervals (CI), comparing crash prevalence for drivers with and without ADHD. Models accounted for multiple crashes per driver using an independent correlation structure and adjusted for potential confounding factors selected a priori, including age at licensure, sex, race/ethnicity, insurance payor, census tract-level median household income and population density, diagnosis of anxiety disorder or DBD, and year of birth. For the comparison of at-fault crashes, we tested for statistical interaction of ADHD with sex using type 3 score statistics. Although we did not conduct formal tests of interaction of other crash outcomes by sex given limited power due to the rarity of some outcomes, we also present sex-specific proportions for completeness and consider these exploratory analyses. Finally, we conducted several sensitivity analyses restricted to: (1) each driver’s first crash within 48 months of licensure; and (2) all crashes and at-fault crashes that occurred within 12 months of licensure. In accordance with strong guidance from the fields of epidemiology and statistics, we do not conduct null hypothesis significance testing using an arbitrary alpha level; instead, we present interval estimation to convey the precision of point estimates for adjusted models [20]. All analyses were conducted using SAS software, Version 9.4 (SAS Institute Inc., Cary, NC). This study was approved by the CHOP Institutional Review Board.
Results
Table 1 shows demographic characteristics among crash-involved ADHD and non-ADHD drivers. The majority of the cohort were long-term primary care patients, with a median of 19 (interquartile range (IQR): 10, 31) primary care visits and a median age at last CHOP visit of 18.0 (IQR: 16.5, 19.0). Those with ADHD were more likely to be male, non-Hispanic White, and have private health insurance. Most crash-involved drivers were followed for 24 months post-licensure (90.9% of ADHD vs. 92.2% of non-ADHD group) and over two-thirds were followed for 48 months post-licensure (70.6% of ADHD vs. 74.6% of non-ADHD group).
Table 1.
Demographic characteristics of drivers involved in at least one crash within 48 months of licensure, by ADHD status.
| Overall N=6,092 | ADHD N=934 | No ADHD N=5,158 | P-value | |
|---|---|---|---|---|
| Age at licensure, median (IQR), years | 17.1 (17.0, 17.6) | 17.4 (17.0, 18.0) | 17.1 (17.0, 17.6) | <0.001 |
| Age at last primary care visit, median (IQR), years | 18.0 (16.5, 19.0) | 18.2 (17.1, 19.3) | 18.0 (16.4, 18.9) | <0.001 |
| Number of CHOP primary care visits, median (IQR) | 19 (10, 31) | 24 (14, 40) | 18(10, 30) | <0.001 |
| Sex, N (%) | ||||
| Male | 3,062 (50.3) | 677 (72.5) | 2,385 (46.2) | <0.001 |
| Female | 3,030 (49.7) | 257 (27.5) | 2,773 (53.8) | |
| Race/ethnicity, N (%) | ||||
| Non-Hispanic White | 3,862 (63.4) | 702 (75.2) | 3,160 (61.3) | <0.001 |
| Non-Hispanic Black | 980 (16.1) | 112 (12.0) | 868 (16.8) | |
| Non-Hispanic other or unknown race | 1,054 (17.3) | 98 (10.5) | 956 (18.5) | |
| Hispanic | 196 (3.2) | 22 (2.4) | 174 (3.4) | |
| Insurance payor, N (%) | ||||
| Private | 5,683 (93.3) | 900 (96.4) | 4,783 (92.7) | <0.001 |
| Medicaid/Self-pay | 161 (2.6) | 19 (2.0) | 142 (2.8) | |
| Not recorded or not billed | 248 (4.1) | 15 (1.6) | 233 (4.5) | |
| Anxiety disorder, N (%) | <0.001 | |||
| No | 5,592 (91.8) | 780 (83.5) | 4,812 (93.3) | |
| Yes | 500 (8.2) | 154 (16.5) | 346 (6.7) | |
| Disruptive behavior disorder, N (%) | <0.001 | |||
| No | 5,785 (95.0) | 765 (81.9) | 5,020 (97.3) | |
| Yes | 307 (5.0) | 169 (18.1) | 138 (2.7) | |
| Quintile of median household income of residential census tract, $, N (%) | 0.17 | |||
| 1st: < 48,506 | 398 (6.5) | 57 (6.1) | 341 (6.6) | |
| 2nd: 48,506 – 66,937 | 1,245 (20.4) | 169 (18.1) | 1,076 (20.9) | |
| 3rd: 66,938 – 84,921 | 2,081 (34.2) | 323 (34.6) | 1,758 (34.1) | |
| 4th: 84,922 – 110,035 | 1,827 (30.0) | 288 (30.8) | 1,539 (29.8) | |
| 5th: ≥ 110,036 | 540 (8.9) | 97 (10.4) | 443 (8.6) | |
| Unknown | 1 (0.0) | 0 (0.0) | 1 (0.0) | |
| Quintile of population density of residential census tract, population per square mile, N (%) | 0.31 | |||
| 1st: < 1,299 | 2,305 (37.8) | 370 (39.6) | 1,935 (37.5) | |
| 2nd: 1,299 – 2,919 | 1,919 (31.5) | 279 (29.9) | 1,640 (31.8) | |
| 3rd: 2,920 – 5,237 | 1,348 (22.1) | 215 (23.0) | 1,133 (22.0) | |
| 4th and 5th: ≥ 5,238 | 519 (8.5) | 70 (7.5) | 449 (8.7) | |
| Unknown | 1 (0.0) | 0 (0.0) | 1 (0.0) | |
| Number of crashes as driver within 48 months | <0.001 | |||
| 1 | 4,635 (76.1) | 661 (70.8) | 3,974 (77.0) | |
| 2 | 1,110 (18.2) | 194 (20.8) | 916 (17.8) | |
| 3+ | 347 (5.7) | 79 (8.5) | 268 (5.2) |
Of these crash-involved drivers, those with ADHD were more likely to crash multiple times. Over the first 48 months of licensure, 934 drivers with ADHD were involved in 1,308 police-reported crashes, while 5,158 drivers without ADHD were involved in a total of 6,676 crashes. By 48-months post-licensure, 29.2% of drivers with ADHD and 23.0% without ADHD had experienced ≥2 crashes (Table 1). Within 24 months of licensure, 12.2% of drivers with ADHD and 9.3% of those without were involved in 2 crashes, while 2.6% and 1.6%, respectively, were involved in 3 or more crashes (χ2 test p=0.007) yielding a total of 874 crashes among drivers with ADHD and 4,501 crashes among drivers without ADHD.
Drivers with ADHD were determined to be at fault for a higher proportion of their crashes than their non-ADHD counterparts within 24 months (77.5% vs 69.4%, aPR: 1.09 [1.04, 1.14]) and 48 months (74.3% vs 66.7%, aPR: 1.09 [1.05, 1.14]) of licensure. Table 2 shows the most frequent crash-contributing driver actions for crashes occurring within 24 and 48 months of licensure by ADHD status. For both drivers with and without ADHD, the most common crash-contributing actions was inattention, followed by unsafe speed, failure to yield the right of way, following too closely, and backing unsafely. Crash-involved drivers with ADHD were more likely to have been inattentive compared to those without ADHD (aPR: 1.10 [1.01, 1.20] within 24 months, aPR: 1.15 [1.07, 1.23] within 48 months); point estimates suggest they may also have higher rates of crashes due to unsafe speed (48 months: 10.7% vs. 8.2%, aPR: 1.10 [0.90, 1.35]), but CIs were notably wide given the relative rarity of speed-related crashes.
Table 2.
Proportion of most frequent crash-contributing driver actions among crash-involved drivers, and proportion of most frequent crash types among at-fault crash-involved drivers, with adjusted prevalence ratios (95% CI), within 24 and 48 months of licensure, by ADHD status.
| All crashes within 24 months | All crashes within 48 months | |||||
|---|---|---|---|---|---|---|
|
|
|
|||||
| ADHD n (%) | No ADHD n (%) | aPR (95% CI)b | ADHD n (%) | No ADHD n (%) | aPR (95% CI)a | |
|
| ||||||
| Total number of crash-involved drivers | 874 | 4,501 | 1,308 | 6,676 | ||
| Driver actions | ||||||
| At-fault | 677 (77.5) | 3,122 (69.4) | 1.09 (1.04, 1.14) | 972 (74.3) | 4,451 (66.7) | 1.09 (1.05, 1.14) |
| Inattention | 395 (45.2) | 1,833 (40.7) | 1.10 (1.01, 1.20) | 592 (45.3) | 2,625 (39.3) | 1.15 (1.07, 1.23) |
| Unsafe speed | 97 (11.1) | 380 (8.4) | 1.12 (0.88, 1.42) | 140 (10.7) | 548 (8.2) | 1.10 (0.90, 1.35) |
| Failed to yield right of way | 87 (10.0) | 438 (9.7) | 1.08 (0.85, 1.37) | 111 (8.5) | 598 (9.0) | 0.96 (0.78, 1.19) |
| Following too closely | 85 (9.7) | 380 (8.4) | 1.02 (0.81, 1.29) | 130 (9.9) | 596 (8.9) | 1.00 (0.83, 1.20) |
| Backing unsafely | 52 (5.9) | 238 (5.3) | 1.16 (0.85, 1.57) | 64 (4.9) | 310 (4.6) | 1.10 (0.83, 1.47) |
|
| ||||||
| Total number of at-fault crash-involved drivers | 667 | 3,122 | 972 | 4,451 | ||
| Crash type among at-fault drivers | ||||||
| Rear-end (index [striking] vehicle) | 232 (34.3) | 1,063 (34.0) | 1.01 (0.89, 1.14) | 364 (37.4) | 1,565 (35.2) | 1.05 (0.95, 1.15) |
| With non-motor vehicle | 130 (19.2) | 511 (16.4) | 1.07 (0.89, 1.30) | 185 (19.0) | 719 (16.2) | 1.11 (0.95, 1.30) |
| Right-angle | 107 (15.8) | 531 (17.0) | 0.95 (0.77, 1.18) | 139 (14.3) | 723 (16.2) | 0.91 (0.76, 1.09) |
| Side-swipe (same direction) | 61 (9.0) | 261 (8.4) | 1.05 (0.79, 1.40) | 87 (9.0) | 401 (9.0) | 0.97 (0.77, 1.23) |
| Struck parked vehicle | 45 (6.6) | 230 (7.4) | 1.00 (0.72, 1.40) | 55 (5.7) | 312 (7.0) | 0.90 (0.67, 1.22) |
| Backing | 41 (6.1) | 226 (7.2) | 0.88 (0.63, 1.24) | 53 (5.5) | 309 (6.9) | 0.85 (0.63, 1.15) |
| Left turn/U-turn | 26 (3.8) | 155 (5.0) | 0.73 (0.48, 1.11) | 30 (3.1) | 204 (4.6) | 0.64 (0.44, 0.95) |
Adjusted prevalence ratio and 95% confidence interval were obtained using generalized estimating equation log-binomial regression model, accounting for within-driver correlation and controlling for age at licensure, sex, race/ethnicity, insurance payor, census tract-level median household income and population density, diagnosis of anxiety disorder or DBD, and year of birth.
Crash types among at-fault drivers are shown in Table 2. During the first 24 months of licensure, more than two-thirds of at-fault drivers with ADHD crashed in one of three ways: (1) as the index (striking) vehicle in a rear-end crash (34.3%); (2) crashing with a non-motor vehicle (19.2%); and (3) a right-angle crash (15.8%). Distributions were similar for adolescents without ADHD. These three crash scenarios remained the most common types of at-fault crashes for drivers with and without ADHD over the first 48 months of licensure. One notable difference was that point estimates suggest drivers with ADHD were an estimated 36% less likely than drivers without ADHD to crash while making a left- or U-turn (48 months of licensure: 3.1% vs. 4.6%; aPR=0.64 [0.44, 0.95]).
We found no evidence of multiplicative interaction in the likelihood of a driver being at fault for their crash by ADHD status and sex within 24 months (p=0.08) or 48 months (p=0.20). Figure 2 displays exploratory analyses of sex-stratified comparisons of outcomes. For both females and males, the likelihood of being at fault for crashes 48 months post-licensure was higher for drivers with ADHD, as was the proportion of drivers who were inattentive. Notably, while males with and without ADHD had similar rates of unsafe speed (10.8% vs. 10.3% of crashes, respectively), the proportion of females with ADHD whose crashes were due to unsafe speed was over 1.5 times higher than crashes among females without ADHD (10.3% vs. 6.4%). With respect to crash types, sex-specific trends were similar to overall trends; one notable exception was the increased likelihood among crash-involved drivers with ADHD of crashing with a non-motor vehicle was limited to females (males with vs. without ADHD: 18.8% vs. 18.9%; females: 19.6% vs. 13.7%) (Figure 3).
Figure 2.
Proportion of most frequent crash-contributing driver actions among crash-involved drivers within 48 months of licensure, by ADHD status and sex.
Figure 3.
Proportion of most frequent crash types among at-fault crash-involved drivers within 48 months of licensure, by ADHD status and sex.
In sensitivity analysis, point estimates for all and at-fault crashes within 12 months of licensure were generally similar with wider confidence intervals. Additionally, restricting to a driver’s first crash within 48 months showed similar results as all crashes within 48 months.
Discussion
This is the first study to compare real-world crash circumstances for a large sample of adolescent drivers with and without ADHD who were involved in crashes in the first 48 months of licensure. Crash-involved drivers with ADHD were an estimated 9% more likely to be at-fault for their crash and have a higher proportion of their crashes attributed to driver actions, such as inattention. Additionally, with the exception that drivers with ADHD were less likely to crash while making a left- or U-turn, drivers with and without ADHD who were at-fault for their crash experienced different types of crashes in similar proportions. Analyses also suggest female drivers with ADHD may have a higher risk of colliding with a non-motor vehicle and crashing due to unsafe speed than females without ADHD; there were no differences in circumstances among males.
Recent research established adolescent drivers with ADHD have somewhat higher crash risk than their peers without ADHD [3,8,12]. Our finding that drivers with ADHD are more likely to have inattention noted by police as a crash-contributing action is aligned with knowledge of the core symptoms of ADHD (i.e., inattention, hyperactivity, and impulsivity). Overall, however, drivers with ADHD appear to crash in similar ways as their peers without ADHD; this is incongruent with naturalistic and simulator studies that report drivers with ADHD are more likely to make various driving errors—e.g., improper turns, following too closely—and have higher rates of elevated gravitational-force driving events [10,11]. Thus, based on our initial findings, we hypothesize the differences in proximate factors that contribute to the heightened crash risk of drivers with ADHD are nuanced, and therefore may need to be observed in more depth than what is possible via crash reports.
One way to better elucidate these nuances may be by conducting prospective studies that utilize naturalistic driving methodologies to follow adolescents with ADHD during early independent driving. Such studies will also be able to more comprehensively characterize driving behaviors, uncover specific proximate factors and real-world driving contexts that place teens with ADHD at increased crash risk, and directly inform how driver training may be best adapted for adolescent drivers with ADHD. Naturalistic driving studies in early licensure will also be able to provide more direct insights on why we found certain differences in crash circumstances between females but not males (e.g., higher risk for ADHD group in crashing with an object). Community sample studies of children with ADHD have suggested symptom profiles may differ in ways that might affect driving. For example, boys may experience greater hyperactivity, impulsivity, and externalizing problems while girls more internalizing problems (e.g., anxiety, depression) and potentially higher levels of inattention [21–24]. Future research should examine males and females separately to uncover potential differences in the mechanisms underlying increased crash risk.
While more work is still needed to better characterize why drivers with ADHD are more likely to crash than their peers without ADHD, our results and prior research showing that teens with ADHD have higher rates of moving violations for engaging in unsafe driving behaviors (e.g., speeding, seat belt nonuse, electronic equipment use) suggest the immediate need for risk-mitigating interventions [12]. Until tailored interventions for drivers with ADHD are developed, there is some evidence to support the following approaches to reducing crash risk. First, prior studies have found that ADHD medication reduces crash risk among drivers with ADHD [8]; however, few adolescents with ADHD are medicated at the time of licensure [3], and adherence with medication among adolescents with ADHD has been found to be poor [25,26]. Thus, families should work closely with their providers to bolster new driver’s medication and adherence plans when appropriate. In addition, supervised practice driving in the learner period that focuses on skills highlighted as crash-contributing factors like attention, speed management, hazard awareness and perception training, and right of way can help with skill development [27]. Finally, we encourage the use of certified driving rehabilitation specialists, who are trained to assess the strengths and weaknesses of drivers with neurodevelopmental or learning differences and can tailor the learning-to-drive process appropriately [28].
Strengths of this study include its longitudinal nature and increased generalizability to the general population of adolescents with ADHD by using a community-identified cohort. A primary limitation is that given our sample includes individuals who were diagnosed prior to adolescence, some may not have impairments due to ADHD symptoms by the time of licensure [29,30]; thus, our results more appropriately represent individuals with a lifetime history of ADHD. The prevalence of ever-diagnosed ADHD in our study cohort (14%) was somewhat higher than a recent US-based national estimate of 10% [2], perhaps reflecting parental decisions to seek care at CHOP. In addition, as NJ has the oldest licensing age (17) in the US and is highly urbanized, results may be less generalizable to US states with much lower minimum driving ages and with highly rural populations. Crash-contributing driver actions were determined and recorded by the responding police officer, which may have introduced subjectivity into assessment of driver actions and may depend on the driver’s willingness to disclose behaviors such as ‘inattention’. However, NJ publishes a field guide to provide specific instructions to law enforcement to support systematic and accurate completion of data fields on the NJ Police Crash Investigation Report [15]. Finally, data limitations precluded us from determining ADHD medication status at the time of the crash, which may have affected crash circumstances (e.g., decreased severity); future prospective studies should be designed to assess medication status among crash-involved drivers.
With notable exceptions, adolescent drivers with and without ADHD had generally similar distributions of crash types. Future naturalistic driving studies are needed to identify more clearly the most important proximate (i.e., in-vehicle) factors that underlie observed elevated crash rates among young drivers with ADHD.
Implications and Contribution.
This study compares crash-related circumstances among young drivers with and without ADHD, highlighting the need for additional support in driving safety (e.g. mitigating inattention, medication management) and future in-depth studies to identify key pre-crash factors and mechanisms underlying elevated crash rates for drivers with ADHD.
Acknowledgements:
The authors would like to thank Haley Bishop, Miriam Monahan, and Rachel Myers for their review of this manuscript. We would also like to thank Melissa Pfeiffer for her data linkage and data management work on the NJ-SHO data warehouse.
Sources of Funding:
This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development at the National Institutes of Health Awards R01HD079398 and R01HD096221 (PI: Curry). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The sponsor had no role in the: design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
List of Abbreviations:
- ADHD
Attention-deficit/hyperactivity disorder
- CI
confidence intervals
- CHOP
Children’s Hospital of Philadelphia
- EHR
electronic health record
- ICD-9-CM
International Classification of Diseases, Ninth Revision, Clinical Modification
- NJ
New Jersey
- PR
prevalence ratios
Footnotes
Disclosure of Potential Conflicts of Interest Statement: Dr. Yerys has received funding from Aevi Genomic Medicine and F. Hoffman-La Roche Ltd. in the past three years. None of the other authors have no biomedical financial interests or potential conflicts of interest to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- [1].Danielson ML, Bitsko RH, Ghandour RM, et al. Prevalence of Parent-Reported ADHD Diagnosis and Associated Treatment Among U.S. Children and Adolescents, 2016. J Clin Child Adolesc Psychol 2018;47:199–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Xu G, Strathearn L, Liu B, et al. Twenty-Year Trends in Diagnosed Attention-Deficit/Hyperactivity Disorder Among US Children and Adolescents, 1997–2016. JAMA Netw Open 2018;1:e181471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Curry AE, Metzger KB, Pfeiffer MR, et al. Motor Vehicle Crash Risk Among Adolescents and Young Adults with Attention-Deficit/Hyperactivity Disorder. JAMA Pediatr 2017;171:756–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Fischer M, Barkley RA, Smallish L, et al. Hyperactive children as young adults: driving abilities, safe driving behavior, and adverse driving outcomes. Accid Anal Prev 2007;39:94–105. [DOI] [PubMed] [Google Scholar]
- [5].Reimer B, Mehler B, D’Ambrosio LA, et al. The impact of distractions on young adult drivers with attention deficit hyperactivity disorder (ADHD). Accid Anal Prev 2010;42:842–51. [DOI] [PubMed] [Google Scholar]
- [6].Barkley RA. Behavioral inhibition, sustained attention, and executive functions: constructing a unifying theory of ADHD. Psychol Bull 1997;121:65–94. [DOI] [PubMed] [Google Scholar]
- [7].Curry AE, Power TJ. Editorial: Paving the Way Toward Improving Safety Among Drivers With Attention-Deficit/Hyperactivity Disorder. J Am Acad Child Adolesc Psychiatry 2020;59:923–925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Chang Z, Quinn PD, Hur K, et al. Association between medication use for attention-deficit/hyperactivity disorder and risk of motor vehicle crashes. JAMA Psychiatry 2017;74:597–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Bernstein J, Roye S, Calamia M, et al. Utility of a novel simulator paradigm in the assessment of driving ability in individuals with and without attention-deficit hyperactivity disorder. ADHD Atten Deficit Hyperact Disord 2019;11:403–11. [DOI] [PubMed] [Google Scholar]
- [10].Merkel RL, Nichols JQ, Fellers JC, et al. Comparison of On-Road Driving Between Young Adults With and Without ADHD. J Atten Disord 2016;20:260–9. [DOI] [PubMed] [Google Scholar]
- [11].Klauer C, Ollendick, Ankem G, et al. Improving Driving Safety for Teenagers with Attention Deficit and Hyperactivity Disorder (ADHD). National Surface Transportatoin Safety Center for Excellence; 2017. [Google Scholar]
- [12].Curry AE, Yerys BE, Metzger KB, et al. Traffic crashes, violations, and suspensions among young drivers with ADHD. Pediatrics 2019;143:e20182305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Curry AE, Pfeiffer MR, Metzger KB, et al. Development of the integrated New Jersey Safety and Health Outcomes (NJ-SHO) data warehouse: catalysing advancements in injury prevention research. Inj Prev 2021;27:472–478. [DOI] [PubMed] [Google Scholar]
- [14].Gruschow SM, Yerys BE, Power TJ, et al. Validation of the use of electronic health records for classification of ADHD status. J Atten Disord 2019;23:1647–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].State of New Jersey Motor Vehicle Commission. New Jersey NJTR-1 Crash Report Manual 2017. [Google Scholar]
- [16].Curry AE, Pfeiffer MR, Myers RK, et al. Statistical implications of using moving violations to determine crash responsibility in young driver crashes. Accid Anal Prev 2014;65:28–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Af Wåhlberg AE, Dorn L. Culpable versus non-culpable traffic accidents: what is wrong with this picture? J Safety Res 2007;38:453–9. [DOI] [PubMed] [Google Scholar]
- [18].United States Census Bureau. Census Gazetteer files, 2010. Available at: http://www.census.gov/geo/maps-data/data/gazetteer2010.html. Accessed April 5, 2018.
- [19].United States Census Bureau. 2013–2017 American Community Survey 5-year estimates. Available at: https://www.census.gov/newsroom/press-kits/2018/acs-5year.html. Accessed April 15, 2019.
- [20].Wasserstein RL, Lazar NA. The ASA’s Statement on p-Values: Context, Process, and Purpose. Am Stat 2016;70:129–33. [Google Scholar]
- [21].Dupaul GJ, Jitendra AK, Tresco KE, et al. Children with attention deficit hyperactivity disorder: are there gender differences in school functioning? School Psych Rev 2006;35:292–308. [Google Scholar]
- [22].Gaub M, Carlson CL. Gender differences in ADHD: a meta-analysis and critical review. J Am Acad Child Adolesc Psychiatry 1997;36:1036–45. [DOI] [PubMed] [Google Scholar]
- [23].Gershon J. A meta-analytic review of gender differences in ADHD. J Atten Disord 2002;5:143–54. [DOI] [PubMed] [Google Scholar]
- [24].Soffer SL, Mautone JA, Power TJ. Understanding girls with attention-deficit/hyperactivity disorder (ADHD): Applying research to clinical practice. Int J Behav Consult Ther 2008;4:14–29. [Google Scholar]
- [25].Sibley MH, Kuriyan AB, Evans SW, et al. Pharmacological and psychosocial treatments for adolescents with ADHD: An updated systematic review of the literature. Clin Psychol Rev 2014;34:218–32. [DOI] [PubMed] [Google Scholar]
- [26].Molina BSG, Hinshaw SP, Swanson JM, et al. The MTA at 8 years: Prospective follow-up of children treated for combined-type ADHD in a multisite study. J Am Acad Child Adolesc Psychiatry 2009;48:484–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Thomas FD, Rilea SL, Blomberg RD, Peck RC, Korbelak KT. Evaluation of the safety benefits of the risk awareness and perception training program for novice teen drivers (Report No. DOT HS 812 235). Washington, DC: National Highway Traffic Safety Administration; 2016. [Google Scholar]
- [28].Monahan M, Classen S. Best practices in driver education to enhance participation. In: Frolek Clark G, Rioux JE, Chandler B, editors. Best Practices for Occupational Therapy in School, 2019, p. 437–45. [Google Scholar]
- [29].Cherkasova M, Sulla EM, Dalena KL, et al. Developmental course of attention deficit hyperactivity disorder and its predictors. J Can Acad Child Adolesc Psychiatry 2013;22:47–54. [PMC free article] [PubMed] [Google Scholar]
- [30].Roy A, Garner AA, Epstein JN, et al. Effects of Childhood and Adult Persistent Attention-Deficit/Hyperactivity Disorder on Risk of Motor Vehicle Crashes: Results From the Multimodal Treatment Study of Children With Attention Deficit and Hyperactivity Disorder. J Am Acad Child Adolesc Psychiatry 2020;59:952–963. [DOI] [PMC free article] [PubMed] [Google Scholar]