Distracted Driving Laws and Motor Vehicle Crash Fatalities

This is a retrospective, time series analysis of MVC fatalities in novice teenage drivers 16 to 19 years old. The study was deemed exempt from review from our institutional review board.

We obtained counts of all motor vehicle fatalities involving drivers and occupants 16 to 19 years old from passenger cars, light pickup trucks, vans, and sport utility vehicles from the Fatality Analysis Reporting System. The Fatality Analysis Reporting System contains data collected by the National Highway Traffic Safety Administration on all MVCs on public roadways in the United States resulting in a death within 30 days of the crash.²³  This database contains detailed crash-related information on the vehicles, drivers, occupants, and circumstances. Counts of fatal crashes involving 40- to 55-year-old drivers were also obtained as a comparison group for the analyses. We used US Census Bureau data for the state mid-year age-specific population estimates to calculate population-based rates of fatal MVCs for each year per age group.²⁴ 

We collected data on individual state motor vehicle laws from the Governor’s Highway Safety Association, Insurance Institute for Highway Safety, and previous studies in which researchers documented dates of legislative enactment (Table 1).^25–27  Laws were considered active only if they were in effect for the entire year of the study. Distracted driving laws were categorized by 3 primary types as defined by the Governor’s Highway Safety Association: (1) Texting laws explicitly banning text messaging while driving. These laws were further coded as being primarily enforced (law enforcement able to pull over and issue citation for text messaging alone) or secondarily enforced (law enforcement can only cite drivers who are pulled over for another offense). (2) Handheld cellphone bans in which all handheld mobile electronic devices are prohibited from use while driving for drivers of all ages, but hands-free device use is allowed. Although law language is highly variable by state, we assumed these laws to include prohibition of texting as it would require use of the handheld device. (3) All-cellphone-use bans that prohibit the use of all hands-free and handheld devices only for novice drivers. Novice drivers were defined differently across all states, with the least restrictive applying only to drivers in their first year of licensure and the most restrictive applying to all drivers <21 years old. As of the end date of our study, no states banned all cellphone use for all drivers (handheld and hands-free) (Table 1, Fig 1). Arizona and Missouri have laws banning only text messaging while driving in only novice drivers; Therefore, they were coded as not having a primary or secondary texting ban and as having no handheld cellphone ban, as defined by our study.

We calculated unadjusted MVC fatality rates for crashes involving the combined 16- to 19-year-old age group of drivers as well as for each individual year of age (16, 17, 18, 19 years old). Additionally, we calculated MVC fatality rates for a comparison group of adult drivers 40 to 55 years old to compare the teenage group to a more-experienced driving population who may be less susceptible to distracted driving by cellphones.¹⁶  Next, we developed 4 age-specific regression models to estimate fatal MVC incident rate ratios (IRRs) with 95% confidence intervals (CIs) for crashes involving the 16-, 17-, 18-, and 19-year-old drivers. We compared these age-specific IRRs under each type of distracted driving law (Table 2) with states not having the specified law, accounting for the specific year(s) a law may or may not have been enacted. Additionally, we conducted the same analyses for the combined 16- to 19-year-old and 40- to 55-year-old driver age groups. We developed a separate multivariate model for crashes resulting in the fatality of a 16- to 19-year-old passenger.

The natural logarithm of age-specific state population was used as an offset term in the models. Our data showed overdispersion and a conditional variance higher than the conditional mean, and thus we used a negative binomial regression model. In the adjusted regression models to control for possible confounding from differences in other state-level factors, we included other motor vehicle traffic safety laws.^9,11,28  These indicator variables based on previous literature included the following: annual changes made to the following motor vehicle safety–related laws, rural interstate speed limits (≤65 or >65 miles per hour), primary enforcement seatbelt laws, immediate administrative license revocation (ALR) for driving with a blood alcohol concentration that exceeds the legal limit, and GDL laws.²⁹  The GDL laws were labeled as “weak,” “moderate,” or “strong” on the basis of categorization by the Insurance Institute for Highway Safety.²⁵  To account for possible differences in macroeconomic factors and crude changes in driving patterns among states, we also included in the multivariate model annual unemployment rates (linear annual unemployment rate for each state) and annual motor fuel usage (a linear term to adjust for annual state-specific highway motor fuel use).^30,31  We used a pooled longitudinal analysis whereby each state contributed 1 observation per year for a total of 550 observations.⁷  Because all covariates were measured annually, they also varied within states over time. A P ≤ .05 was considered statistically significant. All analyses were conducted by using SAS version 9.2 (SAS Institute, Inc, Cary, NC).

Between 2007 and 2017, there were 38 215 fatal MVCs involving 16- to 19-year-old drivers, for an overall fatal MVC rate of 20.1 out of 100 000. The mean fatal MVC rate for the comparison adult drivers was 18.2 out of 100 000 40- to 55-year-old persons between 2007 and 2017. Over the study period, the rate of MVCs involving 16- to 19-year-old drivers combined decreased from 29.5 out of 100 000 persons in 2007 to 18.7 out of 100 000 persons in 2017. For 40- to 55-year-old drivers, the rate of fatal MVCs decreased from 21.2 out of 100 000 persons in 2007 to 19.3 out of 100 000 persons in 2017. In 2007, only 15 states had any form of distracted driving law (1 primary enforcement texting ban, 10 all-cellphone-use bans for novice drivers, and 4 handheld cellphone bans for all drivers) (Fig 1). By 2017, 40 states had primary enforcement texting bans, 6 states had secondary enforcement texting bans, 34 states banned all cellphone use for novice drivers, and 12 banned handheld cellphones for all drivers (Table 1, Fig 1).

Fatal crashes involving teenage drivers increased with each year of age, from a low of 10.7 out of 100 000 persons 16 years old to a high of 27.2 out of 100 000 persons 19 years old (Table 2). When comparing by type of distracted driving law, the unadjusted MVC fatality rate for 16- to 19-year-olds combined was lowest in states with primarily enforced texting bans at 15.9 out of 100 000 persons. It was higher in states with secondarily enforced texting bans at 19.2 out of 100 000 persons and the highest in states with no texting bans at 24.5 out of 100 000 persons (Table 2).

In age-specific, adjusted models when compared with states with no texting law, states with primarily enforced texting bans were associated with decreased incidence rates of fatal MVCs involving drivers of all age groups, with 16-year-old teenagers demonstrating the lowest rates (adjusted incidence rate ratio [aIRR]: 0.64; 95% CI: 0.57–0.71) (Table 2). This association was also observed in adults 40 to 55 years old but to a lesser degree (aIRR: 0.88; 95% CI: 0.83–0.93). Secondarily enforced texting laws were also associated with lower fatal crash rates involving 16-, 17-, and 19-year-old drivers when compared with states with no texting bans, with no association seen in 18-year-olds or the combined 40- to 55-year-old driver age groups.

Similar to the texting ban laws, states with handheld device ban laws for all drivers compared with states with no handheld device ban laws were associated with lower MVC fatality rates in all combined and age-specific models. As with the texting ban laws, the 16-year-old age group demonstrated the greatest reduction (aIRR: 0.57; 95% CI: 0.49–0.67). Of note, novice driver all-cellphone-use bans (n = 276 state-years, Table 1) showed no association with fatal crashes in any age group.

Among teenage passengers 16 to 19 years old, there were 10 300 MVC fatalities in the United States between 2007 and 2017. Primarily enforced texting laws were associated with a 38% reduction in passenger fatalities in teenagers 16 to 19 years old (aIRR: 0.62; 95% CI: 0.57–0.67). Secondarily enforced texting laws were associated with a 27% reduction in teenage passenger fatalities (aIRR: 0.73; 95% CI: 0.64–0.83), and handheld cellphone bans in all drivers were associated with a 25% reduction (IRR: 0.75; 95% CI: 0.67–0.84) (Table 3). There was no association between novice driver all-cellphone-use bans and passenger fatalities in this age group.

In our study, it is suggested that implementation of distracted driving laws over the past 11 years has been associated with a lower rate of fatal MVCs involving teenage drivers in the United States as well as lower teenage passenger fatalities. Over the study period, the number of states with any type of distracted driving law increased from 15 in 2007 to 47 in 2017 (Fig 1). During this time, rates of fatal MVCs involving 16- to 19-year-old drivers decreased by nearly one-third. Among crashes involving the combined group of 16- to 19-year-old drivers, primary enforcement texting laws were associated with a 29% lower adjusted incidence rate of fatalities. Secondary enforcement texting laws and handheld bans for all drivers were associated with a 20% lower adjusted MVC fatality rate. For specific age groups, all distracted driving laws were associated with a lower incidence of fatal crashes, except for novice driver all-cellphone bans.

Our findings are consistent with previous studies in which researchers examined the effects of texting and handheld cellphone ban legislation on fatal MVCs in adults. Self-reported cellphone use in adolescents has been found to be lower after handheld device bans³²  as well as in adults.^33–35  In many of these studies, however, researchers did not examine if decreased cellphone use was due to an increased use in using cellphones as texting devices rather than as handheld phones. An early national study of the impact of texting bans found those applying to all drivers were associated with decreases in single-vehicle, single-occupant car crashes.³⁶  Another study demonstrated primary enforcement texting laws were associated with greater reductions in MVC fatalities than secondary enforcement laws, particularly in younger drivers ages 15 to 21. This study also concluded handheld bans were most effective in drivers 22 to 64.²¹  Our study is the first to analyze each teenage driving age group separately and found a stronger association with handheld cellphone bans in 16- and 17-year-olds compared with other driver ages, although all age groups saw a reduction in association with universal handheld cellphone bans. Each of these different laws has inherent challenges in enforcement. Texting bans are particularly difficult to enforce because law enforcement are often unable to ascertain whether a driver was texting or using a phone for other purposes (eg, for directions).^37,38  Nevertheless, the very presence of such laws may independently deter teenage drivers from the behavior.

In other studies, researchers have found less-promising results from distracted driving laws. A North Carolina study examining the effect of a cellphone restriction on teenagers found no effect on rates of cellphone use 5 months and 2 years after implementation.^39,40  Interestingly, there was a 24% decrease in the likelihood drivers were observed holding a phone to their ear but a 39% increase in the likelihood drivers were observed physically manipulating a phone.⁴⁰  As texting has increased in popularity, more states have responded by implementing universal texting bans. This is supported in our study by the finding that 30 out of 50 states implemented primary enforcement texting laws in 2010 or afterward. In addition to laws, automobile manufacturers, the technology industry, and insurers also have contributed to applications and incentives, which makes it more difficult and less desirable to use a mobile device while driving.⁴¹  Our results provide the most recent and robust data on the longer-term effect of more recently implemented texting laws and their effects on fatal crashes involving teenage drivers and passengers.

Among other distracted driving law analyses, little data exist on the outcome of teenage passenger fatalities in states with varying distracted driving laws. It has been well documented that the presence of passengers can have an effect on driver crash risk, prompting passenger restrictions to be a part of most GDL laws.⁴²  However, it is unknown whether specific distracted driving laws affecting drivers reduce the risk of death in passengers. The largest mortality benefit among passengers in our study was seen in 18- and 19-year-olds in states with primarily enforced texting laws. There was overall a lower incidence of fatal injuries in passengers of all age groups in states with texting laws and universal handheld device bans. These results have important implications for future legislation and the aim of reducing overall MVC-related fatalities in the United States.

As with other traffic law analyses, our study has limitations. We only included MVC fatalities, and not nonfatal MVCs or injuries, because no state-specific database exists with complete nonfatal crash data. Fatal crashes represent the most severe end of the spectrum and are likely to underestimate the effects of traffic safety laws on overall crash risk. No publicly accessible national database of licensed drivers in each state is available. Crash rates differ substantially on the basis of different age combinations and denominators, and we were unable to account for driver experience among these novice drivers.⁴  Population as a denominator allows for changes in overall exposure and has been used in other analyses of novice drivers.⁷  We were unable to assess enforcement of the distracted driving laws or diversion of resources away from other public safety issues as a result of their enactment. As an ecologic study, conclusions about causality cannot be definitively made, but associations can be reported. Although our models were adjusted for additional changes in socioeconomic measures and other driving policies, the findings could be explained by additional factors that were not measured, such as changes in teenage alcohol consumption, numbers of teenage drivers, and technological improvements in automobile safety.

In the United States, primarily enforced distracted driving laws are associated with lower fatality rates in crashes involving 16- to 19-year-old drivers and in teenage passengers. Handheld cellphone bans were effective in all age groups, with the greatest decrease in fatality rates observed in MVCs involving 16-year-old drivers. Novice driver all-cellphone-use bans were not associated with any decreases in fatal MVCs involving 16- to 19-year-old drivers. As rates of MVC fatalities are rising in teenagers, advocating for and implementing effective primary legislation banning handheld cellphone use for all ages will be an important part of reversing this trend. Additionally, practitioners caring for adolescents should continue to counsel on the dangers of driving while distracted to aid in awareness of this important cause for injury and death.

aIRR

adjusted incidence rate ratio

ALR

administrative license revocation

CI

confidence interval

GDL

graduated driver licensing

IRR

incident rate ratio

MVC

motor vehicle crash