Abstract and Introduction
Abstract
Background The use of a cell phone or communication device while driving is illegal in many jurisdictions, yet evidence evaluating the crash risk associated with cell phone use in naturalistic settings is limited. This article aims to determine whether cell phone use while driving increases motor vehicle crash culpability.
Method Drivers involved in crashes where police reported cell phone use (n = 312) and propensity matched drivers (age, sex, suspect alcohol/drug impairment, crash type, date, time of day, geographical location) without cell phone use (n = 936) were drawn from Insurance Corporation of British Columbia Traffic Accident System data. A standardized scoring tool, modified to account for Canadian driving conditions, was used to determine crash culpability from police reports on all drivers from the crashes. The association between crash culpability and cell phone use was determined, with additional subgroup analyses based on crash severity, driver characteristics and type of licence.
Results A comparison of crashes with vs without cell phones revealed an odds ratio of 1.70 (95% confidence interval 1.22–2.36; P = 0.002). This association was consistent after adjustment for matching variables and other covariates. Subgroup analyses demonstrated an association for male drivers, unimpaired drivers, injured and non-injured drivers, and for drivers aged between 26 and 65 years.
Conclusions Crash culpability was found to be significantly associated with cell phone use by drivers, increasing the odds of a culpable crash by 70% compared with drivers who did not use a cell phone. This increased risk was particularly high for middle-aged drivers.
Introduction
Studies from Canada, Australia and the USA have shown that the self-reported prevalence of cell phone use while driving is on the rise, particularly texting while driving, and that young people are particularly inclined to this. Roadside studies indicate that, on average, 2–5% of drivers were observed using a cell phone behind the wheel, whereas a US national survey found that drivers were talking, on average, 7% of the time while they were driving.
Although crash rates in many jurisdictions remain stable or have declined, the proportion of crashes in which a cell phone has been involved has risen. Determining the costs of cell phone-related crashes remains a challenge; however, one estimate suggests that a national ban on cell phone use while driving in the USA would have prevented ~22% of crashes in 2008. A recent report drawing on trends in fatal collisions and mobile device use estimated that >16 000 additional road fatalities, from 2002 to 2007, occurred in the USA as a result of texting on a mobile device. It is not surprising, then, that policies restricting the use of cell phones and other mobile devices while driving are in place in many jurisdictions across the globe.
These policy efforts have been backed by experimental research examining the effects of cell phone use on driver performance. Three recent meta-analyses have synthesized the copious body of experimental research and have concluded that cell phone use produces significant deficits in reaction times, tracking performance and attention, and that these deficits are equivalent for hand-held and hands-free devices, and across age groups. Simulator studies, largely applying cognitive tests to small groups of drivers, have demonstrated that cell phone use negatively affects reaction times, with deficits that compare with those observed in drivers with a blood alcohol level of 80 mg/100 ml.
However, the most important outcome associated with cell phone use while driving is whether its use leads to an increase in the risk of a crash and, in turn, crash-related injury. Experimental studies, although excellent at identifying the mechanisms that affect performance, are not able to fully mimic the complexity of naturalistic driving conditions that most drivers experience daily. Comparatively, little epidemiological research has examined the relationship between cell phone use and crash risk.
A meta-analysis of epidemiological studies of cell phone use and crash risk identified only six studies, of varying research designs, with a reasonable measure of pre-crash cell phone use, and found that cell phone use increased the odds of a crash by 2.86 [95% confidence interval (CI) 1.72–4.75]. Of these, the level of precision in the assessment of cell phone use at the time of the collision varied from a review of cell phone billing records, police reports and in-car camera to driver self-report.
The two most carefully designed studies used case-crossover designs, where each subject serves as their own control, and researchers compare exposure to a transient risk factor (e.g. cell phone use) at the time the outcome of interest (e.g. a crash) occurred, with exposure during a control time period. Using this design, Redelmeier and Tibshirani and McEvoy et al. found a 4-fold increase in crash risk at times when a cell phone was used. This within-person design has much strength, yet not all confounding can be controlled for in a case-crossover design, particularly differences in driving environment, road type, time of day and weather, between exposure and control periods. These studies also lack a crash-free control group, limiting generalizability, as all drivers must have been in a crash to be included in the study (crash-reporting centre, emergency departments).
Another approach to examine cell phone crashes is through a crash culpability analysis. Culpability analysis draws on in-depth data collected from police or traffic engineering reports to assign responsibility to the drivers involved in a crash. These assessments are based on driver actions, the driving environment, traffic flow, road surface, weather conditions and other indicators. When a suitably 'harsh' method of assigning culpability is used, drivers deemed not culpable or responsible for their crash are likely to have done everything correctly during the driving event, and the crash was because of events that were entirely beyond their control. In this situation, non-culpable drivers are believed to closely approximate crash-free drivers.
The objective of the current study is to estimate the association between cell phone use and driver culpability in a sample of drivers involved in crashes from 2005 to 2008 in British Columbia. The current study builds on the epidemiological research examining the impact of cell phone use on crash risk in naturalistic driving conditions while addressing some of the weaknesses of previous studies.