CHAPTER 1
Background

Fatal crashes involving teen drivers remain a significant concern in transportation safety (Simons-Morton et al., 2015). The primary causes include immaturity, inexperience, and risky behaviors such as speeding, kinematic risky driving (e.g., hard braking or hard cornering), and secondary task engagement (Simons-Morton et al., 2017). Klauer et al. (2014) showed that 16-year-old novice drivers have higher crash/near-crash (CNC) risk when performing secondary tasks than adult drivers engaging in the same activities.

Inattention to the roadway resulting from secondary task engagement is a major contributing factor to crash risk for teen drivers (Simons-Morton et al., 2014; Klauer et al., 2017; Gershon et al., 2019), as well as drivers of all ages (Klauer et al., 2006). The operational definition of secondary task implies attention is being diverted away from the driving task. Eyes off road (EOR) is a crucial mechanism for assessing this inattention or diversion of attention away from the driving task. Long glances away from the roadway (exceeding 2 seconds) have been shown to be associated with CNC occurrence (Kingery et al., 2015; Klauer et al., 2017; Aduen et al., 2020). The longer the duration of glance away from the road, the greater the risk of CNC occurrence. The duration of single longest glance exceeding 2 seconds has also been found to significantly increase crash risk to around 4 times that of an alert teen driver (Simons-Morton et al., 2014).

Olsen et al. (2007) summarized various metrics for assessing eye-glance patterns during secondary tasks. These included the number of glances per task, glance frequency (glances to a particular target), glance duration, and percentage of EOR time. Results showed that 29% of novices had glances of greater than 3 seconds for in-vehicle tasks. Additionally, eye-glance patterns of novice teen drivers improved over 6 months of independent driving. Chan et al. (2010) used glances exceeding 2, 2.5, or 3 seconds total time EOR and average glance durations to measure the difference in attention maintenance between novice and experienced drivers. The results indicated that novice drivers are far more likely than experienced drivers to take extended glances inside the vehicle. The single longest EOR and total EOR metrics were also used to examine the association between eye-glance duration and crash risk among teen drivers (Simons-Morton et al., 2014). Single longest EOR glance provided a more sensitive estimate of crash risk than total time EOR (Gershon et al., 2019).

In recent years, the relationship between eye glances and adverse driving outcomes has received increasing interest. However, whether, or to what extent, eye-glance measures can be used to accurately assess crash risk is complicated (NHTSA, 2013). When considering driving experience, novices had more EOR time during challenging in-vehicle tasks than did experienced drivers (Lee et al., 2006; Chan et al., 2010). In addition, the longest single glance off the forward roadway for novices was initially longer than that of their parents. Novices can improve their eye-glance behavior as they gain driving experience (Lee et al., 2006). Over the first 12 months of driving, these novices reduced the longest single glance duration and total percentage of EOR time to levels similar to their parents (OʼBrien et al., 2016).

Attention-deficit/hyperactivity disorder (ADHD) also contributes to the excessive risk of crashes among teen drivers. ADHD is a prevalent neurodevelopmental disorder in children, characterized by excessive levels of hyperactivity and impulsivity and/or inattention [Centers for Disease Control and Prevention (CDC), 2024]. Data from the 2022 National Survey of Childrenʼs Health indicates that approximately 7 million (11.4%) of children aged 3 to 17 years in the United States have been diagnosed with ADHD (CDC, 2024). Replicated evidence has indicated that the core symptoms of ADHD lead to increased engagement in risky driving behaviors and high rates of adverse driving outcomes as compared to adolescents without ADHD (Fuermaier et al., 2017; Curry et al., 2022). Young drivers with ADHD have a crash risk 1.23 to 4 times higher than their neurotypical peers (Klauer et al., 2017; Curry et al., 2019).

To date, much of the knowledge regarding novice teen drivers with ADHD has been derived from crash database or police crash report studies, self-report interviews, controlled experiments, and driving simulators (Barkley et al., 1996; Barkley et al., 2002; Hosking et al., 2009; Curry et al., 2011; Huang-Pollock et al., 2012; Ferguson, 2013; Grabowski et al., 2013; Curry et al., 2019). While beneficial, these studies have limitations, such as difficulty capturing real-world crash situations, recall bias, inaccurate or underreporting of risky/illegal behaviors, and overestimation of driving performance (Jerome et al., 2006).

The association between ADHD symptoms and adverse driving outcomes has consistently positioned inattention as a primary contributor to accidents and violations (Aduen et al., 2020). Drivers with ADHD may have difficulty in attention regulation, may be more susceptible to distraction, may engage in more non-driving tasks (e.g., listening to the radio, looking at objects outside traffic, or having a conversation), may take longer glances away from the road, and may consequently miss important environmental cues or the behaviors of other road users (Fuermaier et al., 2017; Aduen et al., 2019; Epstein et al., 2022). Experience and training can reduce eye-glance rate and crash risk in teens with ADHD. Teens with ADHD after licensure have a 62% higher crash rate in the first month and an estimated 36% higher crash risk during the first 4 years than those without ADHD, regardless of their age and sex (Curry et al., 2017; Curry et al., 2019). Epstein et al. (2022) demonstrated that teen drivers with ADHD exhibit elevated rates of long glances away from the roadway, which are associated with crash risk, and a training program significantly reduced the frequency of long eye glances (lasting ≥ 2 seconds) in teen drivers with ADHD, producing lower CNC rates.

Notably, although previous studies have associated crash risk with EOR, ADHD status, age, and driving experience, few studies have directly examined the crash risk of the EOR among teen drivers of different ages, levels of experience, and ADHD statuses. Identifying key differences in crash risk of EOR between teens with and without ADHD will greatly improve understanding of the factors contributing to this populationʼs high crash rate. This information will guide the development of intervention strategies to improve the learning-to-drive process for ADHD teens.

Existing naturalistic driving data provide an opportunity to objectively and accurately examine teen driver distraction behavior immediately preceding driving events such as CNCs. The second Strategic Highway Research Program (SHRP 2) Naturalistic Driving Study (NDS) has provided a unique opportunity to make deductions a priori from collected observational driving data regarding secondary tasks performed by drivers of all ages in real time, including teen drivers. Additionally, the Virginia Tech Transportation Institute (VTTI) research team has had access to one other relevant teen driver NDS: the Supervised Practice Driving Study (SPDS) NDS. The SPDS NDS has allowed us to evaluate 90 teen driversʼ data over the first 12 months of independent driving. Stratified sampling techniques have allowed us to estimate both the prevalence of distracted driving behaviors over time and the ways different behaviors contribute to the occurrence of CNC events.

The two datasets have provided valuable insights into the contributing factors to crash risk associated with visual-manual secondary tasks, or those tasks that require both visual and physical (i.e., manual) manipulation (Klauer et al., 2014); eye-glance duration (Simons-Morton et al., 2014); risk and prevalence of secondary tasks by age (Guo et al., 2017); and the types of secondary tasks performed (Guo et al., 2017; Hua et al., 2023). Specifically, Guo et al. (2017) found that visual-manual tasks, such as cell phone texting, cell phone browsing, eating, and/or interaction with in-vehicle displays significantly increased crash risk, primarily for drivers aged 16 to 20 years and drivers aged 21 to 29 years. Odds ratios (ORs) were lower, albeit significant, for many of these tasks for drivers aged 30 to 65 years.

Project Objectives

This project analyzed data from the SHRP 2 NDS and SPDS NDS to accomplish the following primary objectives:

1. Assess the association between duration of EOR, crash risk, and driving phases.
2. Determine how the relationship between the duration of inattention and crash risk changes with age.
3. Compare the crash risk of inattention in teen drivers with and without ADHD.