Driver Distraction: Mechanisms, Evidence, Prevention, and Mitigation

Abstract

In this chapter, the reader is introduced to the topic of driver distraction: its definition and mechanisms; its impact on driving performance and safety; approaches to preventing it; evidence-based injury prevention and mitigation countermeasures; and new frames of reference for conceptualizing distraction as traditional driving functions and tasks become increasingly automated. Some strategies that might be considered by societal stakeholders in setting a coordinated agenda for the management of distracted driving going into the future are also presented. Until all vehicles can safely drive themselves, in all conditions, all of the time, it is unlikely that, for driver distraction, Vision Zero will be achieved. In the meantime, however, there is much that can be done to slow its spread and mitigate it effects.

Introduction

Driving is a complex activity that requires, often simultaneously, the performance of one or more driving functions: route finding, route following, velocity control, collision avoidance, rule compliance, and vehicle monitoring (e.g., of fuel status) (Brown 1986). Despite the complexity of this activity, it is common to see drivers engage simultaneously in a range of other, non-driving, activities that have potential to distract them and compromise the performance of these driving functions.

Driver distraction is one of several mechanisms of driver inattention (Regan et al. 2011; Engström et al. 2013) and there is converging evidence that it is a road safety problem (e.g., Beanland et al. 2013; Oviedo-Trespalacios et al. 2016; Dingus et al. 2016). This chapter provides the reader with a general understanding of driver distraction and how to manage it as a road safety issue. We commence by defining what is meant by “driver distraction” and distinguishing it from other forms of driver inattention.

Driver Distraction: Definition, Mechanisms, and Impacts on Driving Performance

Defining “Driver Distraction”

Distraction has been defined inconsistently in the literature (Regan et al. 2011). This is problematic as, in the absence of a commonly accepted definition that can be operationalized and used to code crash and incident data, the role of distraction as a contributing factor in crashes and incidents will be ambiguous (Beanland et al. 2013) – and may lead to quite different estimates of its contribution to crashes and incidents (Gordon 2009). Inconsistencies in definition also make the comparison of research findings across studies difficult, or impossible (Lee et al. 2009).

Driver distraction and driver inattention are related constructs. Like driver distraction, there has been inconsistency in the literature around the definition of driver inattention, and some diversity in thinking about the relationship between the two constructs (Regan et al. 2011). To this end, Regan et al. (2011) attempted to elucidate the relationship between driver distraction and inattention, in the form of a taxonomy of driver attention. The taxonomy was derived from a review of previous classifications of attentional failures identified as having contributed to crashes in in-depth crash studies (e.g., Treat 1980; Hoel et al. 2010; Van Elslande and Fouquet 2007; Wallén Warner et al. 2008). Regan et al. (2011) defined driver inattention as “insufficient or no attention to activities critical for safe driving” (p. 1780) and proposed that driver inattention is induced by five attentional mechanisms (identified in Table 1 below), one of which they labelled “Driver Diverted Attention,” which is synonymous with driver distraction. (See Regan et al. (2011) for a more detailed description of their taxonomy and Engström et al. (2013) for a description of a very similar taxonomy of driver inattention derived from a first principles review of human attentional theory.)

Table 1 Mechanisms of driver inattention (Source: Regan et al. 2011)

Regan et al. (2011) defined driver diverted attention (i.e., driver distraction) as “the diversion of attention away from activities critical for safe driving toward a competing activity, which may result in insufficient or no attention to activities critical for safe driving” (p. 1776). This definition was modelled on an earlier definition formulated by Lee et al. (2009, p. 34) that was subsequently endorsed by an international group of experts convened by the International Organization for Standardization (ISO): “Driver distraction is the diversion of attention away from activities critical for safe driving toward a competing activity.”

Both of these definitions are widely cited in the international literature and are considered suitable by the authors for framing and interpreting the material reported in this chapter. Both definitions carry with them some assumptions (Regan and Hallett 2011):

• Competing activities can be driving-related (e.g., a flashing low fuel warning light) or non-driving related.

• Driver engagement in competing activities can occur involuntarily or be driver-initiated.

• Competing activities can derive from inside the vehicle or outside of the vehicle.

• Competing activities may derive from unknown sources of distraction internal to the mind, such as when daydreaming.

• Driver engagement in competing activities may interfere with the performance of activities critical for safe driving that can be seen (e.g., a lane excursion) or unseen (e.g., a freeway exit missed).

Engström et al. (2013) characterize “activities critical for safe driving” as “…those activities required for the control of safety margins” (p. 17). These include (p. 17) “activities at all levels that are required to maintain acceptable safety margins, such as maintaining headway, keeping in the lane, visually scanning an intersection for oncoming vehicles, deciding whether to yield and interpreting safety-related traffic signs, but excludes those driving-related activities that are not directly related to safety margin control, such as navigation, route finding and eco-driving.”

While performance of a competing activity may divert attention away from any of the driving functions identified by Brown (1986), it is the impact of this diversion on activities critical for safe driving that has been of most interest to the road safety community – and is the reason why the two distraction definitions described above have been framed in the way they have.

Factors That Trigger Driver Distraction

An episode of driver distraction may be triggered through various mechanisms that have been found to relate to a driver’s state, driver needs, properties of the source of the distraction, internal (to the mind) stimuli that trigger distraction, and a driver’s personality characteristics. These mechanisms can, in turn, be classified broadly as either top-down (voluntary; endogenous) or bottom-up (involuntary; exogenous) mechanisms (e.g., Trick and Enns 2009; Lee et al. 2020).

Various driver states may trigger a diversion of attention, including boredom, sleepiness, or fatigue (e.g., Atchley and Chan 2011), social angst (e.g., fear of missing out; Atchley and Warden 2012), and emotionality (e.g., affective state; Chan and Singhal 2013). Driver needs may also trigger a diversion of attention and include the need to communicate with others (Oviedo-Trespalacios et al. 2020a), to be informed (Engelberg et al. 2015), to be entertained (George et al. 2018; Steinberger et al. 2016), and to satisfy basic biological drives like hunger (Irwin et al. 2015). For example, a biological feeling of hunger may trigger a whole chain of internal thoughts about what a driver would like to eat, where they might find what they want to eat, etc., all of which will distract them. These triggering factors that stem from driver states and needs can be characterized as top-down factors (Trick and Enns 2009).

The physical properties of a source of distraction may themselves become distraction triggering factors from a bottom-up perspective. For example, things that are moving, unusual, attractive, unexpected, threatening, salient, or conspicuous are most likely to entice a diversion of attention away from activities critical for safe driving (Regan et al. 2011). Similarly, internal thoughts or internal stimuli from deep within the mind can trigger distraction in a bottom-up manner (as when daydreaming, mind-wandering, or engaged in task-unrelated thoughts; e.g., Smallwood and Schooler 2006). Finally, personality factors, such as a driver’s willingness to engage in distracting activities (Lerner and Boyd 2005) and whether they are particularly vulnerable to attentional capture (distraction prone; Peña-Suarez et al. 2016), may also act as distraction triggers.

There are, in short, many factors that can trigger driver distraction: that is, trigger a diversion of attention away from activities critical for safe driving toward a competing activity.

Competing Activities and Sources of Distraction

A competing activity can be conceptualized as an action performed by a driver on a source of distraction that competes for attention required for the performance of activities critical for safe driving (Regan et al. 2009); for example, as in dialling (the action) a phone number using a mobile phone (the source of distraction). The source of distraction and the actions performed on it by the driver, together, define a competing activity (Regan et al. 2009).

Regan et al. (2009) reviewed seven research studies (five crash studies and two observational studies) in which driver distraction was cited as a contributing factor. They identified around 60 different sources of distraction that gave rise to competing activities in these studies and distilled them into the following broad categories (Regan and Hallett 2011):

• Objects (e.g., mobile phone, advertising billboard, apple).

• Events (e.g., crash scene, lightning).

• Passengers (e.g., child, adult).

• Other road users (e.g., cyclists, pedestrians, other vehicles).

• Animals (e.g., dog).

• Internal stimuli (e.g., that trigger thoughts or the urge to cough or sneeze).

These sources of distraction will be distracting only if drivers interact with them. Regan et al. (2009) identified 53 separate, although not necessarily mutually exclusive (e.g., answering, drinking, listening) actions, that were performed on the various sources of distraction revealed by their analysis.

A consistent finding in the literature is that around 30% of distraction-related crashes derive from driver engagement with distraction sources outside the vehicle. These include animals, architecture, advertising signage, construction zones/equipment, crash scenes, incidents (e.g., road rage), insects, landmarks, road signs, road users, scenery, other vehicles, and weather (e.g., lightning) (Gordon 2009).

A failure to differentiate between a source of distraction and the actions performed on it by a driver can lead to imprecision in the classification of distraction sources. Regan et al. (2009), for example, noted a tendency in some of the studies they reviewed to confound the reporting of events, objects, and actions as sources of distraction. The following, for example, were reported as sources of driver distraction in some of the studies they reviewed (e.g., Gordon 2005): “automobile mechanical problem,” “trying to find destination/location,” “driver dazzled by sunstrike,” “checking for traffic,” and “police/emergency vehicles”.

There is also some confusion in the literature about whether driver states (e.g., fatigue) are themselves sources of distraction. The following, for example, were reported as distraction sources in one of the studies reviewed by Regan et al. (2009; Glaze and Ellis 2003): “driver fatigue/asleep” and “alcohol and fatigue/sleep.” Driver states, such as being fatigued or intoxicated by alcohol, are not in themselves sources of distraction. Rather, they are biological states that can give rise to inattention in the absence of a competing activity (Regan et al. 2011). In the taxonomy of inattention proposed by Regan et al. (2011) (see Table 1), this mechanism of inattention is referred to as driver restricted attention.

Types of Distraction and Triggered Responses

Types of Distraction

A source of distraction has certain “modal properties” (Hallett et al. 2011) which, along with its other physical properties, the state of the driver, drivers’ needs and their personality characteristics, may also trigger a diversion of attention away from activities critical for safe driving.

It is the modal properties of a source of distraction that have been invoked in the literature to define “types” of distraction. An advertising sign, for example, may induce “visual distraction” if a driver looks at it and “internal distraction” (see below) if s/he thinks about the message(s) it conveys (Regan and Hallett 2011). Types of distraction have been characterized in the literature in two ways. Regan (2010) and Regan and Hallett et al. (2011) differentiate as follows between six types of distraction based on the sensory modality through which the diversion of attention toward a competing activity is initiated:

• Diversion of attention towards things that we see (“visual distraction”).

• Diversion of attention towards things that we hear (“auditory distraction”).

• Diversion of attention towards things that we smell (“olfactory distraction”).

• Diversion of attention towards things that we taste (“gustatory distraction”; e.g., the taste of a rotten piece of apple).

• Diversion of attention towards things that we feel (tactile distraction; e.g., the feel of a hairy spider crawling on one’s leg).

• Diversion of attention towards things that we think about (internal or “cognitive” distraction).

It is more common in the literature, however, for “types” of distraction to be differentiated according to the impact that a competing activity has on activities critical for safe driving (e.g., WHO 2011):

• “Visual distraction” – taking one’s eyes off the road

• “Cognitive distraction” – taking one’s mind off the road

• “Auditory distraction” – taking one’s ears off the road

• “Biomechanical distraction” – taking one’s hand(s) off the steering wheel.

There are, however, problems with this latter way of conceptualizing “types of distraction”: (1) it results in an artificially restricted range of distraction types which have potential to interfere with activities critical for safe driving (i.e., it excludes consideration of tactile, olfactory and gustatory distraction); (2) taking one’s ears off the road is really a by-product of taking one’s mind off the road (e.g., as when failing to hear the sound of an approaching motorcycle when engrossed in a mobile phone conversation), rather than a type of distraction per se; and (3) “biomechanical distraction” is actually a form of bimanual, or structural, interference (Kahneman 1973; McLeod 1977) induced by distraction, not a type of distraction per se.

Triggered Responses

The repertoire of driver actions (e.g., answering, listening, writing) that may be performed on all the sources of distraction known to exist is potentially huge. However, the behavioral effects triggered by these driver actions, that may lead to interference (see below) with activities critical for safe driving, appear finite in number. Hallett et al. (2011) have referred to these behavioral effects as “triggered responses” and have characterized them (for distracted drivers) as follows:

• Eyes off the road – driver takes eyes off activities critical for safe driving.

• Mind off the road – driver takes mind off activities critical for safe driving.

• Ears off the road – driver takes ears off activities critical for safe driving (as a result of having one’s mind off the road).

• Hands or feet off controls – driver takes hands and/or feet off activities critical for safe driving.

Conceptualized this way, a given type of distraction (e.g., visual distraction; as defined by Regan and Hallett 2011) may give rise to one or more of these triggered responses, often simultaneously. For example, visual distraction, such as that deriving from the diversion of attention toward an advertising billboard, may take both a driver’s eyes off the road (as when looking at the billboard) and their mind off the road (when thinking about its contents), and while thinking about its contents, their ears off the road (if they become oblivious to auditory information around them critical for safe driving) .

Interference

Triggered responses created by a driver performing an action, or actions, on a source of distraction will likely interrupt or interfere in some way with the performance of activities critical for safe driving.

Driving is a complex, multitask activity (Regan and Strayer 2014; Lee et al. 2009) and different types of attention are required for the performance of activities critical for safe driving, depending on the moment-to-moment requirements of driving. These may include focussed attention, selective attention, divided attention, sustained attention, and switched attention (Wickens and McCarley 2008). Driving, and specifically activities critical for safe driving, also require for their performance the execution of a range of psychological processes that span all stages of the human information processing chain (Michon 1985): detection, perception, short- and long-term memory, decision-making, and responding. Driving typically involves, at any one time, the concurrent execution of multiple tasks, each involving one or more of these types of attention and human information processes. When attention is diverted toward a competing activity, the triggered responses that it generates may interfere with the performance of any or all of these processes during the time that attention is diverted, and may even continue to interfere with activities critical for safe driving for some time after attention returns back to driving (e.g., Strayer and Fisher 2016).

Generally, the degree of interference generated by a competing activity will be a function of three factors (Wickens 2002, 2005):

• The joint demand of the activities critical for safe driving and the competing activity being performed.

• The degree to which both activities compete for access to common human information processing resources (stages of processing [perceptual-cognitive versus action or early versus late processing], processing codes [verbal versus spatial], perceptual modality [auditory versus vocal], and visual channel [focal versus ambient]).

• The manner in which the driver’s attention is distributed between both activities in order to meet their joint demands, whether it is divided between both activities or is focussed primarily on the competing activity.

The research community is still at an early stage, however, in operationalizing the specific mechanisms of interference brought about by distraction, which are discussed further in this chapter in the section “Evidence Implicating Distraction as a Traffic Safety Problem.” While few of these mechanisms have been operationalized, the impacts that they have on driving performance are better researched and understood. They are discussed in the section “Moderating Factors and Self-Regulation.”

Moderating Factors and Self-Regulation

The impact that the performance of a competing activity has on activities critical for safe driving is not constant. The same competing activity (e.g., talking on a mobile phone) may have different effects on activities critical for safe driving depending on factors such as the characteristics of the driver, the demands of driving, the demands of the competing activity, and the ability of the driver to self-regulate their behavior in the face of, or in anticipation of, distraction (Young et al. 2009). Young et al. have labelled these factors “moderating factors” and distinguish between four such factors.

• Driver characteristics: There are characteristics of the driver which may influence the impact of distraction on activities critical for safe driving – by moderating a driver’s willingness to engage in distracting activities, their ability to divide attention between multiple tasks, and their ability to self-regulate their driving in order to maintain suitable safety margins when distracted (Young et al. 2009, p. 340). These characteristics include driver age, gender, driving experience, driver state (e.g., drowsy, drunk, angry), familiarity with and amount of practice with the competing task, and personality (e.g., the propensity to take risks and succumb to peer pressure) (Huth and Brusque 2014; Oviedo-Trespalacios et al. 2020b).

• Driving task demand: The characteristics of the primary driving task itself may influence, in at least two ways, the impact that a competing activity has on activities critical for safe driving: (a) by increasing or decreasing the driver’s mental workload and, hence, reducing or increasing the amount of cognitive resources available for performance of competing activities and (b) by modifying the probability that the driver will have to react rapidly to an unexpected critical event that can give rise to a collision (Young et al. 2009). These characteristics include traffic conditions, weather conditions, road conditions/design, the number and type of vehicle occupants, the ergonomic quality of vehicle cockpit design, and vehicle speed (Li et al. 2020a; Onate-Vega et al. 2020; Oviedo-Trespalacios et al. 2017a, 2020b).

• Secondary task demand: The demands of the competing activity will also influence the degree to which it interferes with activities critical for safe driving, and hence distracts the driver. Secondary (competing) task characteristics that influence secondary task demand include (a) how similar the task is to driving sub-tasks (e.g., whether it requires visual and/or manual control actions similar to those required for performing activities critical for safe driving), (b) its complexity, (c) whether or not it can be ignored, (d) how predictable it is, (e) how easily it can be adjusted, (f) how easy it is for the task to be interrupted and resumed, and (g) how long it takes to perform the task (Young et al. 2009; Regan et al. 2011; Oviedo-Trespalacios et al. 2020b).

• Self-regulation: Self-regulation, in the distraction context, refers to the ability of a driver to self-regulate their behavior in the face of, or in anticipation of, a competing activity in order to compensate for its potentially adverse effects (Young et al. 2009). Young et al. (2009) suggest that self-regulation can occur at the strategic, tactical, and operational levels of driving control (Michon 1985) – in order to regulate their exposure to competing activities (strategic control), to regulate the timing of their engagement in the competing activity (tactical control), and to control mental resource investment in it (operational control). Examples of self-regulation at each of these levels include turning off a mobile phone before a trip (exposure; strategic control), interrupting speech with a passenger when driving through an intersection (timing of engagement; tactical control), and increasing inter-vehicle headway when engaged in a mobile phone conversation (resource investment; operational control) (Saifuzzaman et al. 2015; Oviedo-Trespalacios et al. 2019a; Li et al. 2019; Chen et al. 2020; Bastos et al. 2020).

Impact on Driving Performance

When a driver diverts attention away from activities critical for safe driving toward a competing activity, this may interfere, through the mechanisms discussed, with the performance of driving activities.

Various driving performance deficits are known to arise when drivers are distracted, for a wide range of competing activities – ones that involve interaction with technologies (e.g., mobile phones, iPods, DVD players, navigation systems, e-mail systems, radios, and CD players) and ones involving performance of everyday activities (e.g., eating, drinking, smoking, reading, writing, reaching for objects, grooming, and conversing with passengers). These performance deficits have been discovered in laboratory studies, driving simulators, and in instrumented vehicles driven along test tracks.

The various driving performance deficits reported vary primarily according to the triggered responses induced by the different types of distraction (i.e., eyes off road, ears off road, mind off road, or hands and/or feet off vehicle controls).

Competing activities that primarily take drivers’ eyes off the road have been found to effect specific aspects of driving performance: the selection of information (e.g., failing to detect relevant information from the roadway; spatially concentrated gaze on the forward road center when eyes are returned to the forward roadway); information processing (e.g., longer reaction times to roadway warnings and braking lead vehicles; change blindness that disrupts the detection of changes in the road scene); and vehicle control (e.g., degraded lane keeping performance; reduced speed; increased following distance). For reviews, see Bayley et al. (2009), Horberry and Edquist (2009), and Bruyas (2013). Generally, delays in event detection are greater for competing activities that are visually distracting than for those that are cognitively distracting (that take one’s mind off the road) (Victor et al. 2009).

Competing activities that primarily take a driver’s mind off the road have also been found to affect specific aspects of driving performance: the selection of information (e.g., spatially concentrated gaze on the forward road center; less attention to peripheral hazards; less checking of rear-view mirrors, speedometer); information processing (e.g., inattention blindness, resulting in the “looked but failed to see” phenomenon; memory loss, resulting in an inability to remember some things that have been seen during a drive); and vehicle control (more hard braking; looking less at traffic lights and missing red lights; more navigation errors; reduced variability in lane keeping performance resulting from gaze concentration; no appreciable impact on following distances; acceptance of shorter gaps when turning across oncoming traffic; small decreases in speed; fewer lane changes; more conflicts with vulnerable road users; more traffic rule violations [speeding; red light running; crossing solid lines]; reduced ability to cope with wind gusts; errors [e.g., stopping at green lights and taking off before lights are green]; reduced scanning of intersection areas to the right; and reduced situation awareness [being less able to identify, locate, and respond to hazardous vehicles and to avoid accidents]). For reviews, see Bayley et al. (2009), Horberry and Edquist (2009), and Bruyas (2013).

The authors are unaware of any experimental research that has isolated the impact on activities critical for safe driving of taking one’s hands and/or feet off vehicle controls when distracted (e.g., when steering with one hand while talking on a handheld phone; when steering the vehicle with both knees, as is sometimes seen in video footage from so-called naturalistic driving studies).

Regan et al. (2011; see also Ranney 2008 and Regan 2010) have noted some difficulties in making sense of specific data deriving from studies of the impact of distraction on driving performance. First, it is difficult to rank competing activities in terms of how more or less distracting they are because of differences across studies in methods, measures, and competing tasks employed. Secondly, it is difficult to judge whether a deficit in driving performance within a study brought about by distraction is acceptable, because there is currently no agreement within the international research community on what is an acceptable level of performance degradation for any given competing activity. Finally, the magnitude, and indeed presence, of any performance decrement will be a function of the various moderating factors discussed previously, especially the amount of freedom drivers have to interact in their own way and time with the competing task. Constraining participants to interact with competing tasks in experimental settings in a manner that they would not in the real world may produce performance deficits that simply do not materialize in the real world.

Perhaps one of the greatest difficulties in interpreting driving performance deficits, as pointed out by Regan et al. (2011; see also Wijayaratna et al. 2019), is in knowing to what extent a given reduction in driving performance (e.g., a 20% increase in lateral lane excursions) translates into increased crash risk. Algorithms that link the two remain to be developed and validated.

This section has focussed on the impact of distraction on driver and driving performance. In the following section, we review what is known about the contribution of driver distraction to crashes and crash risk .

Evidence Implicating Distraction as a Traffic Safety Problem

In this section, we examine the impact of distracted driving on traffic safety. We focus here on two types of studies:

• Crash studies that gather information on the frequency and role of distraction involvement in crashes.

• Crash risk studies that aim to provide information about the increased driving risk posed by driver involvement in a distraction-related activity over and above that of the normal risk posed by driving.

Crash Studies

Crash studies use police crash data, medical crash data (from hospital archives), and safety survey data as their main sources of data (Kweon 2011).

In a review of studies using police records from the United States and New Zealand, driver distraction contributed to 10–12% of crashes, and approximately 20% of these crashes involved driver interaction with technology such as mobile phones (Gordon 2009). The Australian National Crash In-depth Study (ANCIS) revealed that 15.9% of crashes were distraction related (Beanland et al. 2013), most commonly involving in-vehicle distraction (13.9%) such as talking with passengers or using the mobile phone. In the USA, a more recent study using the Fatal Accident Reporting System (FARS) database found that 7.7% (13,707 out of 178,677) of all fatal crashes involved distraction (Qin et al. 2019). In a study from Norway, including data from the Norwegian Public Roads Administration (NPRA), it was reported that mobile phones are involved in 2–4% of all fatal crashes, while other in-vehicle distractions excluding mobile phones (i.e., GPS, laptop or tablet computer, video camera, backing camera, passengers, etc.) contributed to 8% of all fatal crashes (Sundfør et al. 2019).

Eby and Kostyniuk (2003) found that rear-end crashes and single-vehicle-run-off-the-road crashes are the two most common types of crash associated with driver distraction. Concerning rear-end crashes, it was estimated that distraction accounts for 21% of all rear-end crashes when the lead vehicle was moving and 24% of all rear-end crashes when the lead vehicle was stopped. Regarding single-vehicle-run-off-the-road crashes, it is estimated that distraction might be the cause of 12–14% of these events.

These studies confirm that distraction is a contributing factor to road crashes. The findings derived from them, however, have some limitations and, as such, must be interpreted with some caution.

Generally, police and hospital crash reports are prone to underreporting of non-fatal cases and a lack of behavioral detail preceding the crash. The lack of behavioral detail around driver distraction could result in an overestimation or underestimation of the problem. In addition, it also limits our capacity to understand the impact of specific behaviors or interactions on crash counts. For example, a common reporting issue in the USA is that a large proportion of crashes reported to involve distraction do not have a specific competing activity listed; rather they specify “distraction/inattention details unknown” (NHTSA 2016). This means that we are often unable to understand the role that technology plays in crash causation in comparison to non-technology distraction or external distractions. Therefore, it is reasonable to argue that crash data should not be the only source of information used for informing evidence-led initiatives for managing distracted driving. More research and innovative data collection and analysis tools are needed to understand the full impact of distracted driving in road crashes.

An emerging alternative to overcome these limitations is the use of naturalistic driving studies, where vehicles are instrumented with video and other sensors to measure driver behavior and performance over extended periods of time. An example of this is the US Second Strategic Highway Research Program Naturalistic Driving Study (SHRP 2 NDS; Dingus et al. 2016), which is the largest naturalistic study ever conducted. The SHRP 2 NDS, also mentioned later on in this chapter, recorded a total of 905 injury and property damage crashes. Dingus et al. (2016) found that observable distractions were associated with 68.3% of all crashes. Given that naturalistic driving studies show the causal link between distraction and crash outcomes (i.e., injury and property damage), at least for observable distraction, it is not surprising that distraction was found to be a greater contributing factor to road crashes in the SHRP 2 NDS than in official records (i.e., police and medical crash data).

Crash Risk Studies

Analyzing crash risk requires additional, supplementary, data on distraction exposure (Kweon 2011); that is, the amount of time spent performing different distraction-related activities while driving. This type of information is typically not collected by police or recorded in hospital archives. Usually, it is collected through safety surveys and in on-road observational studies. It is beyond the scope of this chapter to review in detail all of the literature pertaining to the impact of driver distraction on driver safety. Other resources exist for this purpose (e.g., Cunningham et al. 2017a; Dingus et al. 2016). Rather, we present here an overview of key developments in the understanding of the impact of distracted driving on crash risk.

On-road studies, on which we focus here, comprise naturalistic and quasi-naturalistic approaches that allow for the observation of driver behavior in uncontrolled, or controlled, environments, respectively. In these studies, drivers are observed in their natural driving environment, for weeks or even years, using instrumented vehicles, usually owned by drivers themselves, equipped with video, accelerometers, and other sensors and recording devices (Regan et al. 2013). With new technological developments in in-vehicle driver monitoring, the outcomes of so-called “naturalistic driving studies” (Klauer et al. 2011), which are conducted in uncontrolled environments, are being increasingly reported in the road safety literature. These studies utilize epidemiological methods to sample and analyze the data recorded and provide insightful indications of changes in exposure and risk associated with driver engagement in distracting activities.

Impact of Distraction on Crash Risk

The largest and most comprehensive naturalistic driving study ever undertaken, the US Second Strategic Highway Research Program Naturalistic Driving Study (SHRP 2 NDS; Dingus et al. 2016), involved a comprehensive analysis of the impact of driver distraction on crash risk. Data were collected for 3 years from 3,500 volunteer vehicle drivers, aged between 16 and 98 years. With regards to general distraction (i.e., diverting attention to a secondary task), results from the SHRP 2 NDS demonstrate that, overall, observable distractions increased the odds of having an injury or property damage crash by a factor of 2.0 (odds ratio). An odds ratio (OR) value of 1.0 is considered equivalent to driving while not distracted. Hence, an OR of 2.0 represents a two times increase in crash risk relative to “normal” driving, suggesting that engaging in distracting activities, generally, is a risky activity.

The SHRP 2 study also revealed that, in comparison to other risky behaviors, distraction is one of the most prevalent. Specifically, as can be seen in Table 2, distraction was present during 51.93% of driving time, while other risky behaviors were less prevalent: drug/alcohol impaired driving (0.08%), drowsiness/fatigue (1.57%), speeding (over limit and too fast for conditions; 2.77%), and following a vehicle ahead too closely (0.70%). However, distracted driving risks are relatively lower than some risks generated by other behaviors. Additionally, some distracting activities have been found to be riskier than others. The following section focusses on the risks of some key distracting behaviors reported in the scientific literature, including mobile phone use while driving, and the use of in-vehicle information systems.

Table 2 Crash risk and prevalence of distraction relative to other risky driving behaviors

Impact of Mobile Phones Use While Driving on Crash Risk

Naturalistic studies have provided crash risk estimates for driver engagement in a wide range of secondary activities. In addition to the SHRP 2 NDS, another comprehensive naturalistic study was conducted by Fitch et al. (2013), which aimed to understand handheld and hands-free phone use while driving and its impact on crash risk among 204 drivers during a period of 4 weeks in the USA. In the case of mobile phone use while driving, Fitch et al. (2013) found handheld mobile phone use, overall, to increase the odds of having a crash by a factor of 1.4, while Dingus et al. (2016) reported SHRP 2 NDS data confirming that interaction with a handheld mobile phone, overall, increased the odds of having an injury or property damage crash by a factor of 3.6. These findings, however, can be further considered in terms of the different ways in which drivers use their mobile phones.

The following table illustrates the odds of crash risk associated with driver engagement in specific visual-manual mobile phone tasks while driving. As shown in Table 3, the odds of having a crash increases by 73% for drivers engaged in mobile phone tasks that involve visual-manual interactions (i.e., odds ratio of 1.7). Overall, when considering all of the visual-manual interactions with a handheld mobile phone that have been analyzed while driving as shown in Table 3, “dialling a number on a handheld mobile phone” carries the highest risk (i.e., odds ratio of 12.2).

Table 3 Crash risk associated with visual-manual tasks

Crash risk data for handheld mobile phone conversations and hands-free mobile phone conversations are presented in Table 4. Dingus et al. (2016) reported that handheld mobile phone conversations increase crash risk by more than two times (OR: 2.2; CI:1.6–3.1). In a more recent study, Dingus et al. (2019) found that talking/listening on a hands-free mobile phone did not increase crash risk (did not have an increased OR).

Table 4 Crash risk associated with manual-cognitive tasks (handheld or hands-free device)

Recently, Young (2017) recalculated the odds ratio of handheld mobile phone conversations using the SHRP 2 NDS data after controlling for selection and confounding bias and reported that this resulted in an odds ratio of 0.9. This value is not significantly different from 1, implying that there is no change in risk. It is important to note that this result is similar to findings reported in previous naturalistic studies (Fitch et al. 2013). Table 4 also illustrates the odds of crash risk associated with driver engagement in hands-free mobile phone conversations while driving. As can be seen in Table 3, there is no significant change in crash risk for hands-free conversations, with odds ratios of 0.7 for mobile phone portable hands-free talk and 0.7 for mobile phone integrated hands-free talk (Fitch et al. 2013). Again, neither value is significantly different from 1, further implying that there is no change in risk. Thus, it would seem that handheld mobile phone conversations and hands-free mobile phone conversations are not generally associated with any significant increase in crash risk.

An important warning, however, is necessary here: conversing (speaking or listening) using a handheld or hands-free mobile phone does not occur in isolation in real driving, as implied in the odds ratios reported above. To perform these actions, drivers are often required to first locate the device, reach for the device, dial, or answer the handheld device. These task sub-components of handheld mobile phone conversations could entail highly intensive visual, cognitive, and manual interactions (e.g., dialling or battery/duration monitoring) which could increase crash risk (Oviedo-Trespalacios et al. 2016). For example, in the Dingus et al. (2016) study, reaching for a handheld mobile phone was an extremely risky interaction, specifically increasing the odds of crashing by 4.8 times. This result is concerning given that limited public education has been provided with regards to the increased risk associated with this kind of mobile phone interaction (Oviedo-Trespalacios et al. 2017b).

Impact of In-Vehicle Information Systems (IVIS) on Crash Risk

As the capabilities of in-vehicle information systems (IVIS) have continued to expand over the years, questions have arisen as to whether or not the use of such systems for entertainment (i.e., infotainment) creates risks on the road.

With regards to crash risk, Dingus et al. (2016) found that driver interaction with IVIS increased the odds of having a crash by 4.6 times among drivers in the USA. The same study found this behavior to pose a higher crash risk in comparison to other risky driving behaviors such as fatigued driving (odds ratio = 3.4) and overall handheld mobile phone use (odds ratio = 3.6). A recent meta-analysis conducted by Ziakopoulos et al. (2019), however, found operation of an IVIS to cause only a small percentage of safety-critical incidents, specifically only 1.66% of total crashes. It is important to note, however, that the results from this study were based on a small number of older articles published from 1996 to 2012, when the range of IVIS technologies and functions was more limited. As the capabilities of IVIS continue to increase, more current, up-to-date, research is required to determine the risks associated with use of these systems .

Impact of Interactions with Passengers on Crash Risk

Dingus et al. (2016) also analyzed crash risk associated with active driver interactions with passengers. Crash risk was calculated using data collected from video segments where drivers interacted with adult/teen passengers 6 s prior to a crash. Information related to talking on a handheld mobile phone while driving, discussed above, was gathered in a similar fashion. The results were as follows:

• Drivers interacted with passengers more frequently (14.5% of the total driving time) in comparison to talking on a handheld mobile phone (3.2% of the total driving time).

• However, talking on a handheld mobile phone while driving increased crash risk by 2.2, while interaction with passengers was associated with only a 1.4 increase in crash risk.

Another meta-analysis conducted by Theofilatos et al. (2018) also calculated crash risks associated with passenger interactions. The analysis included a total of seven studies, and the results were as follows:

• 3.55% of crashes were caused by passenger interactions regardless of age.

• 3.85% of crashes were caused by passenger interactions when teen and child passengers were excluded from the analysis.

Recently, Maasalo et al. (2019) examined fatal crash data to determine the crash characteristics and crash risks of drivers with child passengers. The authors found that:

• Female drivers are involved in twice as many fatal crashes with child passengers in comparison to male drivers.

• Drivers with child passengers have a higher tendency to engage in distractions while driving and pose risks particularly around intersections.

• Drivers with child passengers have fewer risk-taking behavior-related fatal crashes (e.g., through speeding) in comparison to drivers with no child passengers.

• Adult passengers lower drivers’ fatal crash risk by helping drivers with child-related tasks.

Collectively, the evidence suggests that primarily cognitive secondary tasks – that take a driver’s mind off the road – are not associated with increased crash risk (increased odds ratios) relative to all driving but are associated with a small but significantly increased odds ratio relative to model driving (i.e., when drivers are alert, attentive, and sober; OR = 1.25, 95% CI [1.01, 1.54]). (Dingus et al. 2019). The effect on crash risk of driver engagement in primarily cognitive secondary tasks is reliably less severe than engagement in tasks that take the driver’s eyes and/or hands away from the driving task (Dingus et al. 2019).

Impact of External Distractions on Crash Risk

As noted earlier, in the section “Competing Activities and Sources of Distraction,” there are many sources of distraction external to the vehicle that have potential to distract drivers. Apart from advertising signs, very little is known about the impact of these on crash risk. Generally, it is known from the work of Dingus et al. (2016) that an extended eye glance duration to an external object increases the odds of having an injury or property damage crash by a factor of 7.1 (OR). Driver interaction with both in-vehicle and external sources of distraction may, therefore, increase crash risk.

Roadside advertising signs are designed deliberately to attract and maintain driver attention to information that is irrelevant to driving. In their meta-analysis of existing studies investigating digital roadside advertising signs (i.e., moving images and/or film clips), Sisiopiku et al. (2015) found an increased crash risk associated with driver interaction with digital roadside advertising signs. However, the effect was only observed on sections of road with intersections. Experimental research, on the other hand, suggests that crash risk increases by approximately 25–29% in the presence of digital roadside advertising signs (Oviedo-Trespalacios et al. 2019b). Fixed object, side swipe, and rear-end crashes have been found to be the most common types of crashes in the presence of roadside advertising signs (Islam 2015; Sisiopiku et al. 2015).

Impact of Other Distractions on Crash Risk

Some other distractions have also been shown to increase the odds of crashing (Dingus et al. 2016):

• Reading and writing (including with tablets) – by 9.9 times.

• Reaching for objects inside the vehicle (excluding mobile phones) – by 9.1 times.

• Drinking (non-alcohol) and eating – by 1.8 times.

• Personal hygiene activities – by 1.4 times.

Generally, as noted above, distracting activities that carry the greatest crash risk are those that involve both visual-manual interactions and occupy a greater proportion of a driver’s time. Particularly troublesome, in this respect, is the use of handheld mobile phones which, in the Dingus et al. (2016) study, increased crash risk overall by 3.6 times and engaged them for 6.4% of their driving time.

Prevention of Distracted Driving

The road system is complex and, from a distraction perspective, many stakeholders are ultimately responsible for preventing and managing distraction (Department of Transport and Main Roads 2020b): drivers, regulatory and enforcement agencies, infrastructure planners, the insurance industry, the mobile connectivity industry, road users and their associations, the automotive industry, technology providers, the telecommunications industry, employers, and the research community.

Traditionally, the onus of responsibility for safe driving has been on the driver. This approach implies that drivers are solely responsible for road safety and thus are to blame for a crash by not following a particular road rule (Newnam and Goode 2015). Generally, this “victim blaming” approach has, to date, been the status quo of distraction prevention. However, safety professionals and academics concur that this approach is unsuitable to deal with distracted driving (or any other risky behavior; Tingvall and Haworth 1999; Tingvall et al. 2009; Young and Salmon 2015).

Drivers tend to be, and will continue to be, distracted due to a number of factors that are often difficult to control. A good example of this is the use of mobile phones, which are a key part of today’s professional and social contexts. Some experts have conceded that ending or reducing phone use is becoming unrealistic (Panova and Carbonell 2018). In addition, there are reports showing that more individuals are establishing maladaptive relationships with their mobile phone, such as “fear of missing out” (FOMO), that could be linked with mobile phone distraction (Elhai et al. 2018; Nguyen-Phuoc et al. 2020). FOMO is a psychological construct that is defined as the persistent desire to stay connected with others’ rewarding experiences and has been linked to both negative affectivity (e.g., stress, depression, anxiety) and increased severity of problematic smartphone use (Wolniewicz et al. 2018). Recent research has shown that problematic mobile phone use, which resembles addiction, is linked with mobile phone use while driving (Oviedo-Trespalacios et al. 2019c). Therefore, if drivers are not able to self-regulate their mobile phone use, it is very unlikely that legal requirements alone will prevent mobile phone use while driving. Several researchers concur that the high prevalence of distracted driving is linked to a heavy focus of legislation on the role of the driver, while ignoring the responsibility of the wider road transport system (Young and Salmon 2015; Parnell et al. 2017; Oviedo-Trespalacios et al. 2019c).

To address these limitations, different philosophies have evolved that recognize that distracted driving is a serious problem with unacceptable consequences (i.e., injuries, economic loss, disruption of the transport system, etc.) that drivers cannot always prevent themselves. Some good examples of alternate philosophies include the Swedish Vision Zero and the chains of responsibility, which are linked to the limitations and capabilities of road users (Tingvall et al. 2009). The common aim of the abovementioned philosophies is to reduce or eliminate the consequences of a road crash. This means that a road transport system assumes variability in human performance and creates safety margins to protect road users from the inevitability of such variability; for example, in the case of vehicles drifting out of their lane, the use of lane departure warnings or roadway tactile edge lining that can alert the driver to potential danger.

The integrated safety chain of responsibility (ISCR) is an approach that has been proposed by Tingvall et al. (2009) in the case of distracted driving (see Fig. 1). The ISCR approach uses the sequence from “normal” driving to a potential crash, broken down in stages of progression towards a crash. These stages are used to identify possible interventions along the chain, such as technology in both the vehicle and infrastructure as well as broader interventions involving police enforcement or community education.

Fig. 1

The integrated safety chain. (Adapted from Tingvall et al. (2009))

The ISCR starts with an understanding of “normal” driving, which includes all the requirements of the driver to achieve this state. Moreover, conceptualizing the notion of normal driving also involves acknowledging that there is a plethora of factors which affect driver performance, such as cognition, motivation, education, police enforcement, and economic incentives. For example, drivers cannot follow a speed limit that they are unaware of or is not appropriately signed. If normal driving is too difficult to achieve, then the number of people capable of driving would be restricted, or we return to blaming the driver for not fulfilling the requirements of normal driving. In the case of distracted driving, evidence around the world shows that a requirement to never be distracted while driving is unrealistic.

The ISCR accepts that deviation from normal driving is going to occur and explains that countermeasures should be applied to correct the deviation back towards normal driving. An example of this, in the case of speeding, is a vehicular speed warning system, referred to as Intelligent Speed Adaptation (ISA; e.g., Regan et al. 2006), which would alert a driver when exceeding the speed limit. If the deviation is not corrected, and the driver finds themselves in an emerging situation, such as being too close to another vehicle or drifting out of the road lane, the ISCR recommends using the vehicle and infrastructure to help the driver regain control. For vehicles drifting out of their lane, the vehicle could automatically take control of the vehicle through electronic stability control (ESC) and lane departure assist systems to return the driver to normal driving or prepare the driver for the potential next phase. If the driver does not regain control, the next phase involves the vehicle preparing for a crash. This could take the form of the vehicle applying automatic braking and traction control. In the final stage, if a crash occurs, both vehicle and infrastructure could help reduce the severity of the consequences with systems such as vehicle airbags and road crash barriers (which serve to attenuate the force of vehicle impacts).

All the different stages from normal driving to a potential crash have the potential to prevent and mitigate the effects of distraction. The ISCR also points to the need to give equal consideration to countermeasures in vehicles as well as road infrastructure. This premise is the basis for the safe system approach which seeks to create a forgiving road environment that allows for driver variability, such as distracted driving. However, an important consideration is that all of these countermeasures need to be rigorously evaluated to prevent unintended consequences or misuse of technology. For example, it has been reported that some drivers potentially stop using their seat belt and start relying on airbags to protect them when their vehicles are fitted with them (Oviedo-Trespalacios and Scott-Parker 2018).

Most recently, prevention approaches for distracted driving have advocated for broadening the scope of intervention beyond a driver-centered approach. The long-established philosophy of the “systems approach,” established by Heinrich (1931), has been proposed to achieve this. The systems approach explains that road accidents and safety (broadly speaking) are emergent properties arising from nonlinear interactions between multiple components across complex sociotechnical systems beyond the immediate road environment. In the case of distracted driving, a systems approach can help in identifying and determining the impact of the wider road system factors that moderate the relationship between distraction and error (Young and Salmon 2015). This approach broadens the scope of the ISCR which focusses on how the immediate road environment can support safe driving behavior and tolerate unsafe behavior (Young and Salmon 2015), without considering the roles that other stakeholders in the distraction ecosystem (mentioned above) have in supporting safe driving. A systems approach responds to the call for a more holistic approach to managing driver distraction, which has traditionally been dominated by a focus on driver behavior change through education and legislation (Tingvall et al. 2009).

A tool for managing road safety following a systems approach is Rasmussen's (1997) Risk Management Framework (RMF). Rasmussen’s RMF is a generic framework that can be used to develop a complete picture of the factors affecting safety in any domain of interest by describing six levels of the system. In the distracted driving domain, the levels have been conceptualized as follows (Young and Salmon 2015):

• Level 1: Government policy and budgeting: At the government level, safety is controlled through the legal system and legislation including the development of behavior-regulating laws and legislation, such as bans; provision of funding for public education; and policy development.

• Level 2: Regulatory bodies and associations: At this level, legislation is interpreted and implemented into rules and regulations (e.g., vehicle design standards). This includes conversion and informing of distracted driving legislation by regulatory bodies, research organizations, and others with a financial interest in distracted driving (such as police and motor vehicle insurers).

• Level 3: Local area government, planning, budgeting: Here, government policy is developed by local councils, including general road rules related to distracted driving. These rules are later implemented in the next two levels.

• Level 4: Technical and operational management: Stakeholders at this level include other influential and authoritative bodies and organizations with a direct influence on distracted driver behavior and decision-making; for example, vehicle and mobile phone manufacturers, the outdoor advertising industry, driver training organizations, road designers, etc.

• Level 5: Physical processes and actor activities: At this level, the focus is on the drivers themselves – the psychosocial influences upon their distracted driving behavior and their actual distracted driving behavior. This level also considers other road users such as passengers, cyclists, pedestrians, etc.

• Level 6: Equipment and surroundings: Here, the focus is on the physical environment and surroundings in which the person drives, including the motor vehicle.

Rasmussen’s RMF posits that safety is maintained through a process called vertical integration, whereby decisions made at the higher levels (i.e., government and regulatory bodies) should influence actions at the lower levels. Likewise, information about the safety performance of the transport system (i.e., driver behavior and crashes) should flow up the hierarchy and influence decision-making at the higher levels (Rasmussen 1997). Consequently, a systems approach to road safety highlights that responsibility for road safety is shared among a broad group of stakeholders, whose decisions and actions interact and affect each other.

The implementation of the systems approach for distracted driving prevention is still in its infancy. Research has highlighted that some groups of stakeholders, directly linked with distracted driving, have not assumed their responsibilities. A good example is the often-complacent roles of mobile phone manufacturers and application developers in the prevention of mobile phone use while driving (Galitz 2017).

An open question on the systems approach is whether or not the current conception of the system has sufficient breadth. As noted previously, interventions to prevent mobile phone distraction while driving have been heavily focussed on the role of the driver, while ignoring the responsibility of the wider road transport or communication authorities (Parnell et al. 2016; Parnell et al. 2017; Young and Salmon 2015). A systemic approach is more likely to succeed in preventing and mitigating the impact of mobile phone use while driving, and this is exemplified in the recent release of Australia’s National Roadmap on Driver Distraction (Department of Transport and Main Roads 2020b) developed by the Queensland Department of Transport and Main Roads in consultation with the Federal Department of Infrastructure, Transport, Cities and Regional Development, along with a wide range of stakeholders (noted above) from industry, academia, and all Australian jurisdictions. The Roadmap was developed through an extensive collaborative design process, with a focus on reducing driver distraction due to mobile devices. The Roadmap contains five overarching strategies to address the challenge of driver distraction: designing for safer interaction; mapping out the adoption of in-vehicle distraction mitigation technology; recognizing the vehicle as a workplace; encouraging greater compliance through enforcement; and changing driver behavior. The Roadmap contains a proposed forward program of work, with a range of projects aligned in support of the five main strategies. The Roadmap is likely one of very few that currently exist that have been developed in a truly collaborative manner involving all key relevant stakeholders in society responsible, directly or indirectly, for the prevention and mitigation of driver distraction.

More recently, it has been suggested that we must also consider the role of other systems, such as the healthcare system, in managing distraction. The link between problematic mobile phone use and mobile phone use while driving might require the use of clinical therapeutical interventions (Oviedo-Trespalacios et al. 2019c).

Consistent with this theme is the Human Factors Integration (HFI) process (e.g., Standards Australia 2016), which requires the specification of human factors requirements that have to be met during all stages of the lifecycle of an engineering product, system, or piece of infrastructure – from concept design through to design, build/implementation, testing, operation (including maintenance), and decommissioning. The purpose of the HFI process is to ensure products and systems are designed from a user-centered perspective to maximize safety, efficiency, user satisfaction, etc. Adherence to an HFI process, in the context of distraction, would ensure that the potential for distraction is considered and mitigated at all stages of the system lifecycle. For example, an engineering consultant, tendering for the design and construction of a new section of roadway, would be required as part of the HFI process to include in the tender a Human Factors Integration Plan that specifies in what ways the piece of road infrastructure will be designed to prevent and mitigate driver distraction during its lifecycle.

Countermeasures for Distracted Driving

Type of Countermeasures

A number of countermeasures have been developed in an attempt to prevent and mitigate distracted driving. However, there is a dearth of evaluations with regards to distracted driving countermeasures. The aim of this chapter is to systematically review countermeasures supported by empirical research, with a focus on those which have successfully reduced the occurrence or impact of distracted driving. Although there are many frameworks that can be utilized to systematically classify the interventions, this chapter will utilize the “Hierarchy of Controls” system, a widely used framework for preventing risks in socio-technical systems.

The hierarchy of controls presents different levels of solutions for the management of identified hazards and risk. In this chapter, we will use the Occupational Health and Safety Assessment Series (OHSAS 18001), which includes five main categories: elimination, substitution, engineering controls, administrative controls, and personal protective equipment (PPE), as can be seen in Fig. 2. The motivation underlying the hierarchy of controls is that more reliable control measures should be utilized rather than measures that are more likely to fail. At the top of the hierarchy is “elimination” which is traditionally considered the most effective countermeasure. Alternatively, countermeasures that rely on individuals behaving in a certain way are considered less reliable. The adaptation of these five categories to consider countermeasures against distracted driving hazards is explained as follows:

• Elimination: Elimination is the first, most effective, control method in the pyramid. With this method, professionals suggest physically removing the hazard completely. While this is the ultimate goal, this method is potentially difficult to implement and not always possible in certain circumstances. This is particularly relevant for distracted driving from mobile phone use, where it has been demonstrated that drivers have difficulties separating from their phones (George et al. 2018). Nonetheless, this method should always be considered first and implemented before the other methods.

• Substitution: If elimination is impossible, the hierarchy of controls recommends moving on to the second category, known as substitution. With this method, hazardous practices/materials are replaced with an alternative, less hazardous, practice or material. This method must also be implemented in the very early phases of development, and it is crucial that the new practice or material either removes or mitigates the hazard in order to be effective. A good example of this is the integration of safer ways of interaction with the mobile phone, using technology such as “workload managers” for distracted driving (NHTSA 2016). Workload managers are driver support systems designed to limit or postpone information that is allowed to come through the phone when the driver’s workload is high, or limit access to complex interactions that it supports. Specifically, when a driver’s workload is high, workload managers can limit or delay information received through their mobile phone, or restrict access to complex interactions facilitated by the devices.

• Engineering controls: If substitution is also not possible, engineering controls are used. These include the modification or addition of physical safety features to the machinery or equipment in order to control identified hazards. For example, a workplace can provide ergonomic chairs to reduce risk of injury to the back and neck or add safeguards to prevent access to dangerous parts of a machine. In the case of distracted driving, engineering controls could involve the use of active safety technologies, such as automatic braking systems to reduce the risk of crashing, or the use of wire-rope barriers to prevent distracted drivers from veering off the roadway or into the path of other vehicles. Blocking mobile phone interactions with applications such as “do not disturb while driving” is another example of an engineering control (see Oviedo-Trespalacios et al. 2019d, for a review of applications to prevent mobile phone distracted driving).

• Administrative controls: The fourth control method in the hierarchy, administrative controls, involves changing the way individuals work through limiting exposure to a hazard. This can be done through installing signs, rotating jobs, etc. The parallel to driving here is driver behavioral interventions including education, legislation and enforcement, and risk awareness campaigns. Some forms of self-regulation, such as only engaging in distractions when the vehicle is stopped and not moving, could also limit exposure to the hazard (Oviedo-Trespalacios et al. 2019a). It should be noted, however, that this method is not always reliable or effective as it is prone to variability of human performance.

• Personal protective equipment: The fifth and final control method is the implementation of personal protective equipment (PPE), such as gloves, safety glasses, earplugs, etc. The rationale behind this method is to protect the body from injury, but it does not eliminate the hazard. Hence, PPE is deemed the least effective. In the case of distracted driving, the corollary to this would be the provision of seatbelts, airbags, and other passive safety elements to minimize the impact of a crash.

Fig. 2

Hierarchy of controls (Adapted from the National Institute of Occupational Safety and Health NIOSH)

Review of Countermeasures

A review of injury countermeasures for distracted driving was undertaken by the authors using the Hierarchy of Controls for distracted driving as an organizing framework. The following table includes the control method used to minimize or eliminate distraction, a description of the method, outcome(s) from its use, and evidence of its effectiveness in achieving the outcome(s). Given that the focus of this chapter is the prevention of distracted driving, personal protective equipment controls were not included because they are post-crash treatments. Prior research studies conducted on post-crash protective measures have concluded that seatbelts, vehicle design, and emergency care can reduce the severity of crashes (Bhattacharyya and Layton 1979).

Elimination

The effectiveness of today’s solutions for preventing distraction-related hazards while driving have been limited. Only fully automated vehicles operating in all conditions all of the time with SAE Level 5 automated driving features follows the elimination principle. When the automation is active, drivers do not have to control the vehicle and therefore vehicle occupants may engage in different activities unrelated to the control of the vehicle. Currently, the availability of fully automated vehicles is limited and restricted to highly controlled environments such as mining sites. It is anticipated that the safety benefits of fully automated vehicles are potentially enormous and would largely eliminate distraction-related hazards (Litman 2020).

Substitution

Regarding substitution, a countermeasure that continues to be suggested is the use of workload managers for driver distraction. Workload managers, as mentioned previously, are designed to minimize distraction by controlling, transforming, or limiting the information flow so drivers can safely manage their driving demands. The NHTSA (2016) considers that minimizing the workload associated with performing secondary tasks with a workload manager will permit drivers to maximize the attention they focus on the primary task of driving. Some of the approaches to achieve this include: (1) simplifying current distractions for more manageable tasks (Oviedo-Trespalacios et al. 2019d) and (2) only allowing drivers to be distracted at points where they can safely resume the driving task (Bowden et al. 2019). Although experimental work has demonstrated that delaying delivery of irrelevant driving-related information to drivers could reduce the impact of distraction (Teh et al. 2018), the technology needs further testing and evaluation.

Engineering Controls: In-Vehicle and Mobile Technology

Engineering controls for vehicles have been developed in increasing numbers in the form of advanced driver support systems (ADAS). ADAS are systems designed to support the driver in their driving task. The logic is that these systems are going to support the driving task by reducing the driver’s demands and includes systems such as semi-automated navigation, blind spot monitors, etc. For partially automated vehicles, there is no evidence that these are going to reduce distraction-related hazards. On the contrary, it is expected that partial automation will result in more distraction due to decreased engagement with the driving task (Cunningham and Regan 2018b; Regan et al. 2020). A study in China with Tesla drivers found that drivers often engage in distracting activities while using the autopilot system (Lin et al. 2018). Similar findings have been reported in the USA, where drivers of vehicles with ADAS, such as adaptive cruise control, report engaging more in mobile phone use and texting (Dunn et al. 2019). Additionally, Matthews et al. (2019) showed that autopilot systems elicit subjective symptoms of fatigue and loss of alertness that last even after the autopilot system has been deactivated. These findings suggest that some ADAS are likely to be facilitating distracted driving.

Another issue raised is that ADAS could increase the likelihood of information overload resulting in distracted driving (Lee et al. 2020). ADAS often use auditory, visual, or a combination of auditory and visual alerts to communicate key information to drivers about the state of the vehicle and to instruct actions. However, there is growing evidence that poorly designed alert systems could increase distraction. An early experiment conducted by Biondi et al. (2014) showed that continuous exposure to auditory stimuli from ADAS negatively affects driving performance. These findings were further confirmed by a naturalistic driving study conducted in Australia, where 34 vehicles were retrofitted with collision avoidance technology which gave audio and visual warnings to drivers. The results showed that, although the system was capable of improving driving behavior, drivers did not want to continue using the system because it was too distracting (Thompson et al. 2018). A study conducted in Spain found that drivers consider GPS navigation, automatic parking systems, and lane departure warnings the most distracting ADAS (Lijarcio et al. 2019). These results highlight the need to further investigate strategies to optimize the role of ADAS as a control to reduce distraction-related hazards.

Applications to reduce mobile phone distracted driving are also engineering controls to prevent distractions. Generally, these applications restrict visual-manual and auditory interactions with the mobile phone while the vehicle is moving. A large number of applications is currently available, with different capabilities and at different stages of maturity (see Oviedo-Trespalacios et al. (2019d) for a comprehensive review of applications). Early findings from studies in Australia and Israel show that using applications aiming to block visual-manual interactions significantly reduces phone pickups and activities such as texting and browsing while driving (Albert and Lotan 2019; Oviedo-Trespalacios et al. 2020a). Nonetheless, reports from users of these applications (e.g., “Do not disturb while driving” for Apple iOS) reveal that the applications do not always stop notifications from instant message applications such as Facebook Messenger and WhatsApp (Oviedo-Trespalacios et al. 2020a). This could have negative implications for road safety given that previous research has found that unexpected incoming notifications are associated with reduced situation awareness while driving (Van Dam et al. 2020). Nonetheless, partially reducing exposure to mobile phone interactions could be a very effective countermeasure option in practice. Unfortunately, surveys in Australia and the USA have concluded that acceptance and adoption of these applications has been low, ranging from 3.8% to 20.5% (Oviedo-Trespalacios et al. 2019e, 2020c, d; Reagan and Cicchino 2018). Further work is needed to increase the effectiveness of this technology in preventing phone use while driving and the uptake of this technology.

Recent developments in in-vehicle technology also include technologies being built into the vehicle with the purpose of reducing distracted driving. A key technology that has been scientifically evaluated is feedback systems. The aim of these systems is to deliver information to drivers about their performance, on the expectation that this information will positively influence their behavior. In an experimental study conducted by Merrikhpour and Donmez (2017), it was found that feedback systems that consider parental norms (i.e., information about a parent’s performance) and real-time feedback (i.e., alarms triggered by long off-road glances) are associated with a smaller duration of off-road glances. Results from these experiments are very promising. Other in-vehicle technology such as in-vehicle interfaces that provide connectivity between smartphones, vehicle displays, and controllers (e.g., Apple CarPlay and Android Auto) has been suggested as a potential countermeasure for distraction. However, currently the potential safety benefits are unknown. Indeed, there is emerging research suggesting that there is risk of distraction from using such technology (Oviedo-Trespalacios et al. 2019f; Strayer et al. 2019; Ramnath et al. 2020).

Engineering Controls: Roads

Road and traffic engineers have considerable scope to manage distraction from some of the sources of distraction deriving from outside the vehicle that were mentioned earlier in this chapter (see section “Impact of External Distractions on Crash Risk”).

PIARC (2016) makes three primary recommendations for preventing serious crashes arising from driver distraction:

• Lower energies through conflict points to within human tolerances – in the event that a distraction-related crash is inevitable, infrastructure measures will generally ensure that vehicle speeds are within human tolerances for serious injury through relevant conflict points.

• Design to provide opportunities for road users to recover from mistakes and noncompliance – e.g., locating crash barriers further from the through traffic lanes provides an opportunity for errant vehicles to recover before hitting the barrier.

• Design to lower the risk of a crash occurring to an “acceptable” level – designing the road to minimize the risk of driver distraction occurring in the first place; e.g., by preventing the road from surprising the road user.

PIARC (2016) recommends the following specific road engineering treatments that can be used to mitigate the effects of driver distraction:

• Hierarchy Level 1 treatments. These include concrete or steel side barriers, wire rope side and median barriers, lateral shift of the road, roundabouts, grade separation at intersections and speed humps.

• Hierarchy Level 2 treatments. These include rumble strips, tactile line markings, speed humps, rough shoulders, and variable speed signs.

The PIARC (2016) document provides high level guidance for road design to prevent and mitigate the effects of distraction. Cunningham et al. (2017b) provide more specific guidance on managing some of the specific external sources of distraction listed earlier in this chapter (see section “Competing Activities and Sources of Distraction”) that traffic engineers have some control over. For example:

• Animals – on road sections where roadway incursions by animals are common and distract drivers, warning signs, and perhaps barriers, can be used to minimize interaction between drivers and animals.

• Scenery – scenic routes and tourist roads are, by definition, distracting and are often located along winding rural roads. Traffic engineers can alert drivers to the potential for distraction along such roads and employ additional engineering control measures to prevent crashes, give drivers more time to recover from the effects of distraction, and reduce impact speeds in the event of a distraction-related crash.

• Architecture – there exist many buildings and monuments that have the potential to distract drivers. It may be possible for engineers to visually mask (e.g., with trees, fencing) prominent architectural structures and features that are known to distract drivers in high-risk locations.

• Crash scenes – so-called “rubber necking” is a common driver behavior around crash scenes. It distracts drivers and cause crashes, and often creates traffic congestion downstream. Possible countermeasures here might include routing traffic away from crash scenes, where possible, and visually masking the scene in some way from approaching traffic.

• Traffic signs – poorly designed traffic signs can, themselves, distract drivers. For example, if they are absent in locations where they should be (e.g., no street name on the road you are turning onto), this may encourage drivers to adopt compensatory search strategies that distract them. Similarly, if signs are poorly designed (e.g., contain too much information or are incomprehensible), this may encourage long eye glances away from the forward roadway. Poorly designed and absent road signs should be avoided.

Ultimately, the road and traffic engineer should strive for a distraction-tolerant road system such that, in the event of a distraction-related crash, no driver or other road user is killed or seriously injured (Tingvall et al. 2009; see also section of this chapter titled “Prevention of Distracted Driving”).

Administrative Controls

Administrative controls cover legislation, authority enforcement of legislation, as well as education programs.

Legislation banning or restricting distraction has been a key control in the prevention of distracted driving (WHO 2011). Some of the legislative approaches target general driver performance, such as “without due care and attention,” which covers a wide range of distracting behaviors. Graduated driver licensing (GDL) is a policy used to keep newly licensed young novice drivers out of harm’s way by restricting driving to times and situations demonstrated to be of lower risk. In some jurisdictions, such as Queensland, Australia, the GDL bans young drivers’ use of hands-free phones or loudspeaker functions while driving (Department of Transport and Main Roads 2020a), both of which are otherwise allowed among fully licensed drivers. Additionally, there is also more specific legislation targeting activities such as talking, text-messaging, or playing video games on handheld mobile phones while driving. Unfortunately, few studies have assessed the impact of legislation on distracted driving, and most of the research has been centered on mobile phone use.

Evaluations of legislation targeting mobile phone use show partial success in preventing this risky driving behavior. A common finding highlighting the positive impact of legislation is the reduction of handheld conversations among drivers after handheld mobile phone bans were implemented in the USA (Rudisill and Zhu 2017; Rudisill et al. 2019b). More recently, an analysis of the 2011–2014 Traffic Safety Culture Index surveys showed that handheld calling bans were associated with fewer calling behaviors overall and in all demographic subgroups. Evaluations of distracted driving legislation in New Zealand (Wilson et al. 2013) and the UK (Johal et al. 2005) have also reported reduced mobile phone use post legislation.

However, in other cases, legislation seems to have had a minimal effect on behaviors, such as texting on a mobile phone. In the USA, for example, Rudisill et al. (2019b) found that universal texting bans were not associated with less distraction. In Europe, Jamson (2013) documented that drivers in the most highly regulated country with respect to mobile phone legislation (Italy) report texting as frequently as those in countries with no legislation. Furthermore, the effect of legislation seems to be heterogenous among different groups of the population. An analysis of research in the USA concluded that phone legislative restrictions have no long-term effect on the prevalence of mobile phone use among novice drivers (Ehsani et al. 2016). Rudisill and Zhu (2017) found that, although there are net reductions in handheld interactions, mobile phone use was higher overall among females, younger age groups, and African Americans.

These mixed results on the effectiveness of the legislation can be partially explained by challenges associated with the enforcement of the legislation. Law compliance frameworks, such as deterrence theory, have shown that drivers are motivated to avoid harmful behavior by fear of negative consequences. Thus, breaking the law is more likely to occur if the swiftness, certainty, and severity of punishment are low (Homel 1988). Thus, sustained police enforcement programs are a key element to guarantee a reduction of distracted driving through legislation.

Studies in the USA have demonstrated that handheld mobile phone bans require robust enforcement to have the desired effect on driver behavior in the long term (McCartt and Hellinga 2007; McCartt et al. 2010). Specifically, high-visibility enforcement programs (i.e., visibility elements and a publicity strategy to educate the public and promote compliance with the law) targeting drivers who use handheld mobile phones have been trialled successfully. In California and Delaware, handheld mobile phone use dropped nearly 33% as a result of high-visibility enforcement (NHTSA 2016). Importantly, there is growing evidence that capacity to enforce mobile phone bans is restricted unless technological and legislative innovations take place. Different evaluations of distracted driving law enforcement have found several important barriers to enforcement of distracted driving legislation (Nevin et al. 2017; Rudisill et al. 2019a):

• Societal factors: Mobile phones often have a utilitarian function in supporting driving, such as through provision of GPS or maps, which makes it difficult to identify distracted driving. Also, mobile phones experience rapid technological change that is often faster than policy cycles.

• Contextual factors: The ability of police to conduct traffic stops safely is often limited and dangerous (i.e., weaving through cars or high-speed traffic).

• Organizational factors: Police functions are diverse, and resources limited, resulting in low prioritization of distracted driving legislation. Additionally, the lack of clear and enforceable polices is also one of the main difficulties: that is, officers cannot always be sure if a driver was texting or using the GPS while enforcing texting bans.

• Interpersonal factors: Many drivers who challenge police officers during traffic stops increase the difficulty of enforcement operations, and there is not sufficient dialogue among police forces regarding distracted driving.

• Individual factors: Police officers largely report engaging in distracted driving themselves and believe that drivers can safely multitask. Thus, the enforcement of distracted driving legislation might be unprioritized and perceived as not legitimate. Also, it was reported that, in many circumstances, detecting distracted driving is difficult without technology.

Another key factor undermining the effectiveness of enforcement operations targeting distracted driving-hazards is behavioral adaptation by drivers aiming to conceal or avoid police enforcement. Drivers have reported scanning the environment, searching for police, covering the phone all the time with their hand, and using the phone on their laps (Oviedo-Trespalacios et al. 2017b). Moreover, Oviedo-Trespalacios (2018) found that drivers who often engage in these behavioral adaptations also report higher engagement in distractions such as texting and browsing on a mobile phone. Alternatively, research has increasingly reported that drivers are using in-vehicle information systems (IVIS) to engage in texting with their mobile phones, making enforcement of mobile phone use legislation more difficult (Oviedo-Trespalacios et al. 2019f). Concerningly, IVIS are distracting even when interfaces such as Apple CarPlay and Android Auto are used (Oviedo-Trespalacios et al. 2019f; Strayer et al. 2019; Ramnath et al. 2020). The fact that drivers are using these behavioral adaptations to avoid police enforcement undermines the effectiveness of this administrative control and must be addressed in the planning of future legislation and enforcement schemes.

The next group of administrative controls reported in the literature are related to Workplace Health and Safety (WHS) controls to prevent mobile phone use while driving. This is a very important group of controls because employment demands have been consistently linked with distracted driving, during both work-related and non work-related driving (Engelberg et al. 2015). Unfortunately, WHS efforts to prevent distracted driving are relatively new and only a few isolated cases have been evaluated. The main work identified confirmed that truck and bus drivers working for organizations that enforced texting bans have lower texting and driving prevalence in comparison to companies without bans (Hickman et al. 2010). Furthermore, additional research on work-related driving has concluded that implementing WHS policies to prevent distracted driving might not be sufficient to prevent this behavior, needing strict enforcement and sanctions to create a safety culture (Swedler et al. 2015a). Truck drivers in Swedler et al. (2015b) study listed the following examples that could be effective in reducing distracted driving:

• Better procedures for communicating with drivers – delivering a noninvasive signal over dispatch devices to indicate that the driver received a message.

• Enforcing bans on distracted driving activities.

• Video-monitoring to observe drivers engaging in distracted driving.

• Monitoring cell-phone usage if driver is using a company-provided phone.

• Locking out devices while vehicle is in motion.

• Automatically updating package delivery drivers’ routes, so drivers do not have to make scheduling/routing decisions while driving.

There is great potential in the WHS space to reduce distracted driving, particularly among people who drive for work. The development of organizational guidelines could provide a great opportunity to increase road safety. Key guidelines to support this process have been developed in Australia by The National Road Safety Partnership Program (NRSPP 2016): “A guide to developing an effective policy for mobile phone use in vehicles.” The process considers elements that can influence distracted driving in organizational settings, such as the current engagement in distracted driving, leadership, education, training, collection, monitoring and analysis of critical incident data, enforcement, mobile phone design, and vehicle purchase and design. There is a need to consolidate and increase the uptake of good road safety practices about distracted driving in the corporate sector.

Education programs have been developed in an effort to reduce and/or prevent drivers from using their mobile phones while driving. A number of interventions have been identified with significant gains in preventing distracted driving. In the USA, the telecommunications company AT&T launched the “It Can Wait” campaign. As part of the program, drivers are encouraged to sign a pledge on their website, encouraging them to make a commitment to never drive distracted (e.g., “I pledge to always drive distraction free.”). In addition, drivers installed an application capable of detecting when a vehicle is moving more than 25 mph and prevent mobile phone notifications. Furthermore, the campaign also launched a virtual reality experience on their website that helps users experience the dangers of distracted driving. As young adults were the primary target audience for this campaign, AT&T started social media campaigns (i.e., “#ItCanWait” hashtag on Twitter) and released a documentary (i.e., “From One Second to the Next”) to raise awareness about dangerous phone use while driving. The campaign evaluation showed a reduction in road crashes and larger awareness of distracted driving risk (Carter 2014). Unfortunately, studies aiming to replicate these results using similar strategies to those in the campaign have not shown the same success. Fournier et al. (2016) reported that neither the distribution of flyers and thumb bands with fear-based slogans (e.g., “It Can Wait”) nor the encouragement of drivers to sign a pledge seemed to reduce overall mobile phone distracted driving. Interestingly, however, the type of mobile phone use behavior did change, as drivers were found to decrease calling behavior but increase texting behavior while behind the wheel (Fournier et al. 2016). This apparent replacement of a risky driving behavior with an even riskier driving behavior highlights the need for more research to investigate the actual effectiveness of this campaign (Fournier et al. 2016).

Some educational campaigns have been aimed at specific groups of the transport system, such as employees (i.e., “It Can Wait” educational program) and parents of young drivers (i.e., “Steering teens safe” educational program). Tailored educational programs involve the use of workshops, lectures, and demonstrations about distracted driving. The “It Can Wait” educational program showed that these activities could be extremely useful in increasing awareness about distracted driving risk and road rules (Hill et al. 2020). In the case of the “Steering teens safe” educational program, parents were trained to use motivational interview frameworks to use with their teens besides being given relevant road safety knowledge so they could improve their communication with their teens (Peek-Asa et al. 2014). Although a reduction of distracted driving behaviors was reported, the success of these educational interventions has been limited. Given the importance of considering the role of additional actors, such as employers, parents, and friends, among others, in the prevention of distracted driving, future developments are needed in this space.

Emergent approaches to education of drivers do not seek to prevent distraction but to upskill drivers to engage safely in distracting behaviors. An innovative example of this is the “FOrward Concentration and Attention Learning (FOCAL)” educational program developed by Unverricht et al. (2019). FOCAL educational training develops the driver’s capacity to self-regulate off-road glances. Experiments conducted after the training confirmed its effectiveness in reducing the severity of distraction. Specifically, drivers who received the FOCAL training engaged in fewer in-vehicle glances longer than 2 s than drivers who received traditional education on distraction-related risks and road rules (Unverricht et al. 2019). Although evidence is limited, and no inferences about crash risks can be reliably made, this is a very innovative approach with the potential of changing the way we train drivers in the future .

Distraction and Vehicle Automation

New technologies are emerging that are capable of supporting and automating many of the functional driving activities performed traditionally by human drivers. These new technologies have been classified by the Society of Automotive Engineers International (SAE International 2018) as falling into two general categories that span six levels of automation:

• Driver support features (also known as advanced driver assistance systems) that provide momentary assistance and warnings (Level 0), steering or brake/acceleration support (Level 1), or both steering and brake/acceleration support combined (Level 2).

• Automated driving features (ADF) that (a) can drive the vehicle under limited conditions but require the driver to either take control when required (Level 3), (b) can drive the vehicle under limited conditions but do not require the driver to take back control (Level 4), and (c) can drive the vehicle under all conditions all of the time without human intervention (Level 5; SAE International 2018).

With increasing technological support and automation, the driving functions and tasks performed by drivers will change, and this will change the repertoire of knowledge, skills, and behaviors required by drivers to maintain safe driving performance (Casner and Hutchins 2019; Fisher et al. 2020; Regan et al. 2020; Spulber 2016). Even now, a modern driver has a unique skill set compared to drivers two or three decades ago; many drivers today have never driven a manual transmission vehicle or have been required to pump their brakes on slippery roads (Spulber 2016). As vehicles become increasingly supportive and automated, so too will the impact that distraction has on activities critical for safe driving. This is because the activities critical for safe driving will themselves change and, ultimately, in vehicles equipped with Level 5 ADFs, there will be no requirement for the driver to perform them at all. But will distraction, as a road safety issue, disappear when there is no requirement for humans to perform any activities critical for safe driving? We briefly explore this and related issues in the sections that follow, drawing on some recent thinking and empirical findings reviewed in Cunningham and Regan (2018a) and Lee et al. (2020) (and see also Kanaan et al. 2020).

Automation Creating Distraction

As vehicles become more automated, the technologies that drive them may, themselves, become a source of distraction for drivers. Evidence already exists showing that automation actions and alerts that are unexpected, because of a lack of training, lack of situational awareness, or some other mechanism, may create “automation surprises” (Hollnagel and Woods 2005), and, in doing so, distract drivers. Even routine alerts and indicators in vehicles equipped with existing driver support features may draw attention away from the road at inopportune moments in time (Lee et al. 2020).

As noted, vehicles equipped with Level 3 automated driving features are classified by the SAE (SAE International 2018) as being capable of driving the vehicle, but only in limited conditions. At this level of automation, the driver is expected to resume control if requested by the vehicle (e.g., if the automation fails or drifts out of its operational design domain). Here, the frame of reference for distraction may become different in the mind of the driver; the requirement to supervise the vehicle automation could itself become a source of driver distraction (Hancock 2009; Lee et al. 2020).

It is well documented that drivers tend to engage in secondary activities when supported by vehicle automation (Lee et al. 2020). Evidence for this has been found both in driving simulators (e.g., Carsten et al. 2012; Jamson et al. 2013) and in instrumented vehicles driven on test tracks (e.g., Llaneras et al. 2013; Dingus et al. 2016). The propensity to do so tends to be greater for technologies that provide higher levels of automation.

More generally, as vehicles become increasingly automated, the role of the driver is expected to shift from being that of an active controller of the vehicle to that of a more passive supervisor of the automated driving system (Desmond and Hancock 2001). There is evidence that this reduction in task engagement can induce “passive fatigue” (reduced attentional capacity arising from driving task demands which are too low (Desmond and Hancock 2001; Saxby et al. 2013) and, in turn, driver inattention (Saxby et al. 2013; Körber et al. 2015). Here, inattention is brought about not by distraction, per se, but by other mechanisms.

Thus, drivers may be distracted either because automation demands their attention at inopportune moments in time or it induces drivers to engage more often and more deeply in non-driving activities (Lee et al. 2020).

Distraction and Takeover Ability

In vehicles equipped with SAE Level 0–2 driver support features, the driver is considered to be driving the vehicle and is supported in performing activities critical for safe driving by a variety of technologies (e.g., Autonomous Emergency Braking; Adaptive Cruise Control). While distraction, when it occurs, may impair the performance of activities critical for safe driving, the technologies themselves may help to mitigate any detrimental impacts this distraction may have (Tingvall et al. 2009), as noted previously.

In vehicles equipped with SAE Level 3 ADFs, in which automation is capable of driving the vehicle in limited conditions, the automation is considered to be driving the vehicle (SAE International 2018). The driver is, however, expected to resume control of the vehicle if requested by the vehicle; e.g., if the automation fails or the vehicle is driven outside of its operation design domain. There is evidence that takeover quality in vehicles equipped with automated driving features is impaired when drivers are distracted (e.g., Merat et al. 2014). Interestingly, however, the speed of the motor actions required to commence the takeover (e.g., to reach for the steering wheel or apply the brakes) appears to be little affected (Zeeb et al. 2015, 2016). Some evidence also exists showing that manual driving performance may be compromised for a considerable period of time after the handover of control to the driver has been completed (e.g., Merat et al. 2014; for reviews, see Cunningham and Regan 2018a and Fisher et al. 2020).

Self-Regulation and Individual Differences

Drivers of manually controlled vehicles often, as noted earlier, self-regulate their behavior in an attempt to manage distraction (e.g., Bastos et al. 2020; Oviedo-Trespalacios et al. 2019a, 2020b; Ortiz-Peregrina et al. 2020; Tivesten and Dozza 2014). There is also some evidence that they self-regulate their behavior in automated vehicles. Jamson et al. (2013), for example, found that drivers supported by automation self-regulated their behavior in conditions of high traffic density in order to reduce the likelihood of them diverting their attention away from the forward roadway.

Large individual differences have been found in the nature and frequency of engagement in secondary activities when driving automated vehicles (e.g., Llaneras et al. 2013; Clark and Feng 2017; see also Fisher et al. 2020). Clark and Feng (2017), for example, investigated the impact of driver age on secondary task engagement during automated driving periods. They found that both younger and older drivers engaged in secondary activities when supported by automation. However, younger drivers mostly used an electronic device, while older drivers mostly conversed. Körber and Bengler (2014) reviewed a number of individual differences that may moderate the involvement and impact of driver distraction in automated vehicles. These include complacency and trust in automation, driver experience, and the propensity to become bored and daydream.

“Vehicle Distraction and Inattention”

Will driver distraction remain an issue in vehicles equipped with SAE Level 4 and 5 automated driving systems that obviate the need at all for a human driver to intervene? After all, in these vehicles, so equipped, there would be no controls, no driver, and the vehicle occupant could simply sit back and let the vehicle do all the driving.

This question highlights again the frame of reference through which distraction is conceptualized. Cunningham and Regan (2018a) speculate that, if there are only a few SAE Level 3, 4, and 5 vehicles in the community fleet, which mix with SAE Level 1 to 2 vehicles, then there may emerge a new frame of reference for distraction. Here, it is possible that drivers of vehicles being operated manually might be distracted by the behavior of vehicles operating autonomously if the latter have been programmed to drive in ways that violate drivers’ expectations; in much the same way that drivers are distracted by the behaviors of others who drive or ride erratically in traffic flows.

If it is the responsibility of vehicles equipped with SAE Level 4 and 5 technologies to perform automatically all activities critical for safe driving, is it possible for such self-driving vehicles themselves to be distracted? Regan (in Lee et al. 2020) has labelled this “vehicle distraction.” Here, again, the frame of reference for conceptualizing distraction will change. But what competing activities, if any, could divert a vehicle’s “attention” (or computational resources), more generally, away from activities critical for safe driving? In fact, what might it mean for a vehicle driving autonomously to be inattentive and, if it was, what might be the mechanisms of inattention? (Cunningham and Regan 2018a, b).

These are interesting questions. For a vehicle to be attentive to activities critical for safe driving, its algorithms will need to be programmed such that the vehicle knows, from moment to moment, to which activities critical for safe driving it should be attending. If so, it will become necessary to specify – a priori – what these activities critical for driving will be, and they will presumably be a subset of the higher-level functional driving activities specified by Brown (1986), referred to earlier. They will change from moment to moment, along any given stretch of road.

But how do software programmers know what activities, critical for safe driving a vehicle operating autonomously, they should pay attention to from moment-to-moment along a stretch of roadway when the research community has not yet itself agreed on what we, as human drivers, should be paying attention to from moment-to-moment (Kircher and Ahlstrom 2016)? For those who have thought deeply about what activities are critical for safe driving (Engstrom et al. 2013; Hallett 2013), we know that this is not a trivial task.

Nevertheless, it is interesting to speculate on by what mechanisms, if any, a vehicle equipped with SAE Level 4/5 automation technology operating autonomously might become inattentive to activities critical for safe driving? The taxonomy of inattention proposed by Regan et al. (2011), noted earlier (see Table 1), is also useful in stimulating thought about the mechanisms by which an SAE Level 5 equipped vehicle might itself become inattentive to activities critical for safe driving. Regan (in Lee et al. 2020; see also Cunningham and Regan 2018a) has speculated on what “vehicle inattention” might mean for each of the five mechanisms of inattention proposed by Regan et al. (2011):

• Driver restricted attention: For a Level 4 or 5 equipped vehicle, this category of inattention might describe a vehicle that “goes to sleep,” so to speak, if, for example, there is a system failure, or some or all vehicle sensors suddenly become incapable of seeing. Here, the vehicle may become inattentive to some or all activities critical for safe driving.

• Driver neglected attention: This category of inattention would seem to be less relevant to the design of vehicles when driven by automation given that, unlike, humans, they will not be prone to the kinds of attentional biases, expectations, and limitations that humans are.

• Driver mis-prioritized attention: For a vehicle driving itself, this category of vehicle inattention might come about if the computer algorithms that drive it fail, through inadequate design, to give attentional priority to the most critical competing activities critical for safe driving at a given moment in time. Even though vehicles when driving themselves will not have the limited attentional capacity of humans, software engineers will nevertheless need to program the vehicle to prioritize who and what the vehicle should pay attention to at any given moment in time during a trip.

• Driver cursory attention: For a vehicle driving itself, this is about not providing enough attention to activities critical for safe driving. Again, this category of inattention may be less relevant to the design of highly automated vehicles given that, unlike humans, they will not have the same limited attentional capacity. Nevertheless, it is incumbent on software engineers to ensure that vehicles driven by automation allocate enough attention (computational resources) to activities critical for safe driving to ensure that they are successfully and safely completed.

• Driver diverted attention: For a self-driving vehicle, this is about distraction. But is it possible for a vehicle operating autonomously to be distracted? It is probable that, if demanded by consumers, vehicle manufacturers may give drivers the option of operating self-driving vehicles manually. Current evidence suggests that there will be some demand from consumers (Cunningham et al. 2019). In this case, it is possible that drivers themselves could become sources of “vehicle distraction.” This might occur, for example, if they attempt to take back control of a fully automated vehicle when they should not, in which case, vehicle “attention” may be diverted by the driver, at least temporarily, away from what the vehicle considers at that point in time to be the activities critical for safe driving to which it must attend. Vehicle distraction might also occur if people elect to drive self-driving vehicles manually (if allowed) in a way that violates the pre-programmed expectations of vehicle algorithms in other vehicles that are being controlled by automation. Here, self-driving vehicles might be seen as being distracted by the behaviors of other self-driving vehicles being operated manually.

The whole issue of what distraction and inattention, more broadly, might mean for self-driving vehicles in future is a fascinating one. The different frames of reference through which distraction may be conceptualized makes it highly unlikely that it will ever disappear as a road safety issue. For vehicles with higher levels of automation, then, countermeasure development will need to focus in future on a somewhat different set of distraction-related issues:

• The prevention of automation surprises (for Level 3 ADFs).

• Support for quality takeover and rapid gaining of control by drivers when requested by vehicle automation (for Level 3 ADFs).

• Prevention and mitigation of secondary-task engagement at inappropriate moments in time when automation is engaged (for Level 3 ADFs).

• Prevention and mitigation of passive fatigue induced by low workload during prolonged periods of automation (for Level 3 ADFs).

• The programming of automated driving features in a way that ensures that the vehicles they control do not violate the expectations of other road users (for Levels 3–5 ADFs).

• The prevention of “vehicle distraction and inattention” (for Levels 3–5 ADFs).

Countermeasure development for driver distraction at higher levels of automation is, however, in its infancy. Those countermeasures known to have been proposed have focussed on a limited number of areas: education and training for maintenance of vigilance of the driving environment and for understanding ADAS/ADF modes and vehicle performance (e.g., Casner and Hutchins 2019; Noble et al. 2020; Regan et al. 2020); human-machine interface design to minimize automation surprises and support safe resumption of manual control (e.g., Carsten and Martens 2019; Campbell et al. 2020); human factors considerations around policy and regulation for vehicle automation (Burke 2020); and use of driver state monitoring technologies and driver feedback to detect distraction and reorient driver attention (e.g., Lee et al. 2009; Lenné et al. 2020).

Conclusion and Strategies Moving Forward

In this chapter, we have introduced the reader to the field of driver distraction: its definition and mechanisms, its impact on driving performance and safety, prevention approaches, countermeasures, and new frames of reference for conceptualizing distraction as traditional driving functions become increasingly automated.

The focus of the chapter has been on driver distraction, although we acknowledge that there are other road users vulnerable to the effects of distraction, including bicycle riders and pedestrians. To our knowledge, there has been no systematic attempt to define distraction from their frames of reference and to define and classify the sources and mechanisms of distraction that lead to interference with activities critical for safe riding or walking. Furthermore, relatively little research has been done to understand the impact of distraction on their performance and safety (Oviedo-Trespalacios et al. 2019b). Prevention and mitigation strategies for these road user groups are, hence, at a relatively early stage of maturity.

Just as activities critical for safe driving will continue to change as vehicles become more automated, so too will the sources of distraction drivers interact with that may impair their performance. These include new infotainment and other technologies being built into the vehicle by manufacturers, special interfaces that provide connectivity between smartphones and vehicle displays and controllers (e.g., Apple CarPlay; Android Auto), and portable devices brought into the vehicle, including smartwatches and other wearables. While there is some limited research on the effects on driver behavior and performance of interaction with these devices while driving (Oviedo-Trespalacios et al. 2019f; Strayer et al. 2019; Ramnath et al. 2020), little or nothing is known about their impact as contributing factors to crashes and increased crash risk. Similarly, we know almost nothing about the impact on crashes and crash risk of distraction created by automated driving features.

While the focus of this chapter has been on the negative impacts that distraction may have on driving performance and safety, there is evidence that distraction may in some circumstances enhance driving performance and improve safety – by, for example, counteracting the effects of fatigue (Williamson 2009; see also Olson et al. 2009). However, the specific mechanisms by which this occurs (e.g., through increased arousal; increased vigilance, etc.) have not, to our knowledge, been researched and operationalized. Further research is needed to understand under what conditions, and how, distraction can be used in a positive way to optimize driving performance.

Laws that regulate the use of particular technology devices (e.g., mobile phones, visual display units) are becoming quickly outdated as new technologies and modes of interaction with them emerge. Australia’s National Transport Commission (NTC) has recently advocated a shift away from technology-based road rules towards technology-neutral approaches for regulating driver distraction (National Transport Commission 2019). This approach would provide (p. 8) (a) “a clear list of high-risk behaviours and interactions that drivers must avoid regardless of the technology involved or the source of distraction” and (b) “reduced uncertainty about ‘proper control’ to address both the observable causes and consequences of behaviours and interactions that can impair a driver’s control of a vehicle.” This would seem to be a positive way forward that focuses more on those behavioral interactions known to increase crash risk (e.g., long eye glances away from the forward roadway) rather than on the technologies that induce them, and provides clearer, evidenced-based guidance to enforcement authorities on what constitutes improper control of vehicles being driven by distracted drivers.

In addition to the guidance already provided in this chapter, we provide in Table 5 some general strategies that might be considered by society in setting a coordinated agenda for the management of distracted driving going into the future. They have been categorized under headings that will be more familiar to road transport agencies: data collection and evaluation, education and training, employers, legislation and enforcement, licensing, public education, research, road and traffic engineering and design, roadside advertising, stakeholder consultation, technology design, and vehicle design. These strategies derive from material presented in this chapter, our own thinking and some other sources (Regan et al. 2009; European Commission 2015; NRSPP 2016; PIARC 2016; Imberger et al. 2020; Regan et al. 2020; Department of Transport and Main Roads 2020b). It is our hope that the material presented in this chapter, along with the general strategies outlined in Table 5, will go some way towards informing the future management of distracted driving.

Table 5 Strategies moving forward to manage driver distraction

Until all vehicles can drive themselves, in all conditions, all of the time, it is unlikely that we will achieve Vision Zero for distracted driving, and even then, self-driving vehicles may themselves be vulnerable to its effects. In the meantime, however, there is much that can be done to prevent and mitigate the effects of driver distraction as we strive, collectively, to achieve Vision Zero.

Cross-References

• Automated Vehicles: How Do They Relate to Vision Zero

• Vision Zero: How it All Started

Appendix

A Review of countermeasures against distracted driving hazards following the “Hierarchy of Controls.”

Level of control	Countermeasure	Description	Evidence of Effectiveness
Elimination	Fully Automated Vehicles (FAV) SAE Level 4–5	FAV use sensors and software to drive the vehicle. Drivers are not required to take control or monitor the vehicle.	As drivers are not required to take control, this raises the question as to whether or not drivers can engage in distracted driving (e.g., eating, mobile phone use, etc.) without any repercussions. Automated vehicles are already being used in controlled road networks such as mining sites (Gershgorn 2016). Projected benefits from such vehicles are only likely to be observed in 25–30 years, mostly due to challenges in infrastructure, legislation, etc. (Dia 2015; Clark et al. 2016)
Substitution	Workload managers	Presenting driving and non-driving related information in such a way that road users are not distracted. Workload managers help drivers to focus on driving when it is needed.	Experimental research has shown that the technology could have a positive impact on safety in reducing distraction from ADAS (driving-related information) (Teh et al. 2018). The effectiveness of this approach to manage non-driving related information is unknown.
Engineering controls	Partially Automated Vehicle SAE Level 3 (also known as autopilot)	Partially automated vehicles involve automation of key vehicle control tasks, e.g., lateral and longitudinal vehicle control. However, drivers are supposed to maintain their hands on the steering wheel and be supervising the vehicle. Also, drivers must be available to take over control of the vehicle at all times.	Drivers are likely to engage in distraction while driving partially automated vehicles (Lin et al. 2018), which could result in road crashes.
Engineering controls	Advanced driver-assistance systems (ADAS) including crash warning system [available in SAE Level 1–2 vehicles]	ADAS support drivers with features such as cruise control, blind spot monitors, lane centering, etc. ADAS also include systems that provide warnings for upcoming collisions.	Early warnings helped decrease drivers’ reaction times, thus avoiding more conflicts and collisions (Lee et al. 2002; Bao et al. 2012). However, ADAS might not compensate for the effects of distraction on drivers’ performance (Sieber et al. 2015). ADAS could increase distraction by means of poorly designed warnings/alarms (Biondi et al. 2014; Thompson et al. 2018; Li et al. 2020b). The spread of in-vehicle information systems (IVIS) increases the capabilities and often licit means to engage in non-driving tasks (Oviedo-Trespalacios et al. 2019f). IVIS are often used as an interface for ADAS.
Engineering controls	Blocking technology for mobile phone distractions	There are two main technologies: mobile phone applications and hardware which seek to block mobile phone use while driving (Oviedo-Trespalacios et al. 2019d)	Mobile phone applications and hardware-based technologies to block mobile phone technologies are effective in reducing mobile phone use while driving (Oviedo-Trespalacios et al. 2020a; Albert et al. 2019). However, there are concerns that some applications will not fully prevent high-risk mobile phone interactions. A field study using hardware-based blocking technologies in Australia showed a reduction on mobile phone use while driving (Ponte et al. 2016). Acceptance of blocking technologies is low among drivers which limits adoption of the technology (Oviedo-Trespalacios et al. 2019e, 2020c; Ponte et al. 2016; Reagan and Cicchino 2018).
Engineering controls/administrative controls	Feedback systems [Targeting young drivers]	Feedback systems deliver information to drivers on their performance. Three main different feedback systems have been established (Merrikhpour and Donmez 2017): Social norms feedback that provided a report at the end of each drive on teens’ distracted driving behavior, comparing their distraction engagement to their parents. Post-drive feedback that provided just the report on teens’ distracted driving behavior without information on their parents Real-time feedback in the form of auditory warnings based on eyes-off-road time	Feedback systems that consider parental norms and real-time feedback are associated with a smaller duration of off-road glances, with parental norms feedback outperforming real-time feedback (Merrikhpour and Donmez 2017). Post-drive feedback showed no significant effect (Merrikhpour and Donmez 2017).
Administrative controls	Legislation	Legislation banning the use of mobile phone devices while driving	There are mixed reports about the effectiveness of bans across jurisdictions. In the USA, universal handheld calling bans are associated with lower self-reported handheld conversations for adult drivers (Rudisill et al. 2019b) as well as a reduction in motor vehicle crash (MVC)-related emergency department (ED) visits (Ferdinand et al. 2019). Alternatively, Rudisill et al. (2019b) found that universal texting bans were not associated with lower texting. A US literature review found that nearly none of the restrictions targeting young distracted drivers sustainably prevented mobile phone use while driving (Ehsani et al. 2016). In Europe, Jamson (2013) found that drivers in the most highly regulated country (Italy) reported texting while driving as frequently as those in countries with no legislation.
Administrative controls	Police enforcement	Police operations to increase compliance of legislation banning different forms of mobile phone use while driving.	Reductions of mobile phone use while driving are observed with resource-intensive police operations (McCartt et al. 2010). However, there are doubts of the long-term impact of the bans on safety (McCartt et al. 2014). The effect of police enforcement of mobile phone bans is limited. Research has shown that police officers are unable to correctly enforce legislation in many circumstances due to lack of resources, poor visibility inside vehicle, unenforceable legislation, among others (Rudisill et al. 2019a; Nevin et al. 2017). Additionally, drivers can engage in police-avoidance strategies such as concealing their mobile phone which further reduce effectiveness of police operations (Oviedo-Trespalacios 2018).
Administrative controls	Organizational procedures/ policies [Targeting driving for work]	Implementation of Workplace Health and Safety (WHS) organizational procedures/policies to reduce distracted driving and enhance safety of drivers.	Truck and bus drivers working for organizations that enforced texting bans have lower texting and driving prevalence in comparison to companies without bans (Hickman et al. 2010). Participants in Swedler et al. (2015b) study emphasized the importance of implementing policies that are clearly and strictly enforced. In addition, a participant in the study mentioned that a texting ban would help create an organizational safety culture that does not normalize distracted driving. Truck drivers in Swedler et al. (2015b) study listed the following examples that could be effective in reducing distracted driving: Better procedure for communicating with driver – delivering a noninvasive signal over dispatch device to indicate that the driver received a message Enforcing bans on distracted driving activities Video-monitoring to observe drivers engaging in distracted driving Monitor cell-phone usage if driver is using a company-provided phone Locking out device while vehicle is in motion Automatically updating package delivery drivers’ routes so drivers don’t have to make scheduling/routing decisions while driving
Administrative controls	“It Can Wait” educational program	In 2010, AT&T launched the “It Can Wait” campaign across the USA to educate drivers on the dangers of texting and driving. The campaign involved: Signing an online pledge to encourage drivers to make a commitment to never drive distracted (e.g., “I pledge to always drive distraction free.”) A mobile app called “AT&T DriveMode,” which detects when a vehicle is moving more than 25 miles an hour. If a driver receives a text message or email when the vehicle is in motion, the app is capable of sending an automatic reply to notify them that the user is currently driving (AT&T 2012). A virtual reality experience on their website that helps users experience the dangers of distracted driving	AT&T evaluated the program after its launch, and found strong, positive, relationships between the campaign activities, particularly social media sharing, pledging, and mobile apps, and the projected reduction in crashes that would have taken place across the USA over a 1-year period (Carter 2014).
Administrative controls	“Steering teens safe” educational program	Educational intervention targeting parents of teen drivers. The intervention included: Parents received a workbook that identified 19 safety lessons divided into 4 topics: basic safety principles (including distracted driving); safe driving skills; rural driving; and special situations. Parents were upskilled on motivational interviewing in an initial 45-minute session. Parents received three 30-minute follow-up phone calls to provide additional intervention support	The intervention had a weak but positive effect, enhancing risky driving and parent-teen communication about road safety (Peek-Asa et al. 2014).
Administrative controls	“Just Drive—Take Action Against Distraction” educational program [Targeting driving for work]	An education program designed to increase awareness of the dangers of distracted driving and to encourage employees to be safe and responsible drivers. The program included presentations, group activities, and a “pledge card” to document the planned behavioral changes. The target group were businesses and organizations in San Diego County as part of employee safety and wellness programs.	The program was well received among the participants and resulted in positive changes in short-term intention and medium-term behaviors. Additionally, participants showed increased knowledge regarding distracted driving legislation and risks (Hill et al. 2020).
Administrative controls	“FOrward Concentration and Attention Learning (FOCAL)” educational program	The FOCAL program is a research-led training that consists of three stages: Pretest: Trainees watch four video clips while having to switch between viewing the forward roadway and a map via keyboard. Trainees were asked to find the location of a street while also limiting glances away from the road to less than two seconds. Participants hit the space key on the keyboard to switch between the viewing the forward roadway and the map Training: This stage included feedback, timer, 3-second in-vehicle glance training, and 2-second in-vehicle glance training. Posttest: Trainees finished the program by watching four new video clips and completing the same task as in the pretest.	Driver who received the FOCAL training engaged in fewer in-vehicle glances longer than 2 s by roughly 25 percentage points when compared to the placebo group (Unverricht et al. 2019).