Introduction Alcohol-impaired driving (AID) crashes accounted for 10 511 deaths in the USA in 2018, or 29% of all motor vehicle-related crash deaths. This study describes self-reported AID in the USA during 2014, 2016 and 2018 and determines AID-related demographic and behavioural characteristics.
Methods Data were from the nationally representative Behavioral Risk Factor Surveillance System. Adults were asked ‘During the past 30 days, how many times have you driven when you have had perhaps too much to drink?’ AID prevalence, episode counts and rates per 1000 population were estimated using annualised individual AID episodes and weighted survey population estimates. Results were stratified by characteristics including gender, binge drinking, seatbelt use and healthcare engagement.
Results Nationally, 1.7% of adults engaged in AID during the preceding 30 days in 2014, 2.1% in 2016 and 1.7% in 2018. Estimated annual number of AID episodes varied across year (2014: 111 million, 2016: 186 million, 2018: 147 million) and represented 3.7 million, 4.9 million and 4.0 million adults, respectively. Corresponding yearly episode rates (95% CIs) were 452 (412–492) in 2014, 741 (676–806) in 2016 and 574 (491–657) in 2018 per 1000 population. Among those reporting AID in 2018, 80% were men, 86% reported binge drinking, 47% did not always use seatbelts and 60% saw physicians for routine check-ups within the past year.
Conclusions Although AID episodes declined from 2016 to 2018, AID was still prevalent and more common among men and those who binge drink. Most reporting AID received routine healthcare. Proven AID-reducing strategies exist.
- motor vehicle - occupant
Data availability statement
Data are available in a public, open access repository. Not applicable.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Motor vehicle crashes in the USA are a significant public health issue that causes death and injury, burden health systems and have negative economic impacts. In 2018, traffic crashes on public roadways in the USA caused 36 560 motor vehicle-related deaths1 and an additional 2.7 million non-fatal emergency department visits.2 These statistics include drivers, passengers and non-occupants such as pedestrians and bicyclists. Alcohol-impaired driving (AID) is a major risk factor for traffic crashes. Of the 36 560 motor vehicle crash deaths that occurred in 2018, 29% (n=10 511) involved an alcohol-impaired driver.1 Both the yearly number of deaths and the number that involved an alcohol-impaired driver have either held steady or increased annually from 2014 through 2018,1 3–6 suggesting that a renewed effort to confront and reduce AID is needed.7–9
Efforts to reduce AID in the past have been successful. Between 1982 and 1997, there was a 43% decrease in the proportion of alcohol-impaired drivers involved in fatal crashes.10 This corresponded with a time when many US states implemented laws making it illegal to drive with a blood alcohol concentration of 0.08 g/dL or higher and grassroots organisations like Mothers Against Drunk Driving (MADD) were formed to promote policies to reduce AID.11 Strategies addressing AID have the potential to substantially reduce motor vehicle crashes and deaths.12 Effective strategies to prevent AID exist, including drunk driving laws, sobriety checkpoints, ignition interlocks, mass media campaigns and increasing alcohol taxes.9 13 However, implementation of these strategies varies across states and communities.14–16
The total number of self-reported AID episodes among adults in the USA per year has been estimated to range from 110 to 160 million during 1993 through 2012 with no clear decrease over time.17 18 In 2012, an estimated 1.8% of adults in the USA reported at least one AID episode during the previous 30 days, which translated to 4.2 million adults engaging in 121 million annual AID episodes (a rate of 505 per 1000 population).18 An update to these estimates is needed to illustrate the continued call for universal implementation of prevention efforts using both established and promising strategies.
This study estimated the annual prevalence, number of episodes and rates of AID among adults in the USA during 2014, 2016 and 2018. We also examined how these outcomes varied by certain demographic and behavioural characteristics.
Data were from the 2014, 2016 and 2018 Behavioral Risk Factor Surveillance System (BRFSS) surveys. BRFSS is a nationally representative, cross-sectional, ongoing, random-digit-dialled telephone survey. State health departments in collaboration with the US Centers for Disease Control and Prevention use trained interviewers to collect reported health-related behaviours from a representative sample of civilian, non-institutionalised adults aged ≥18 years residing in any US state or territory. BRFSS participants are recruited via landline and cellular telephone numbers. All BRFSS questionnaires and data are available online.19 Because the BRFSS is a surveillance system, the Centers for Disease Control and Prevention’s Institutional Review Board has determined that the BRFSS is exempt from its review.
Nearly half a million adults completed the interview in each year (456 664 in 2014; 486 303 in 2016 and 437 436 in 2018). We limited the analysis to adults residing in the 50 US states or the District of Columbia that had information recorded for the AID survey question. The median response rates for the19BRFSS 2014, 2016 and 2018 surveys were 47% (49% landline, 41% cell phone), 47% (48% landline, 46% cell phone) and 50% (53% landline, 43% cell phone), respectively.
In even-numbered years, BRFSS respondents who reported having had at least one alcoholic beverage in the past 30 days were asked ‘During the past 30 days, how many times have you driven when you have had perhaps too much to drink?’ Responses were recorded as whole numbers ≥0 and were considered to be the number of AID episodes. Those who reported no alcohol in the past 30 days were coded as having zero AID episodes. We created a binary variable for AID (yes/no) categorising people reporting zero episodes as ‘no’ and those with ≥1 episodes as ‘yes’.
Respondent demographic characteristics collected included age in years at the time of the survey, race and ethnicity, highest level of education obtained, current marital status and household income. Reported behavioural characteristics collected included binge drinking and seatbelt use. Binge drinking was defined as having on at least one occasion five or more drinks for men and four or more drinks for women during the past 30 days. Seatbelt use was ascertained by asking ‘How often do you use seatbelts when you drive or ride in a car? Would you say—always, nearly always, sometimes, seldom or never?’ Responses were categorised into a binary variable: always versus less than always. AID prevalence, episodes and rates were described across demographic and behavioural characteristic categories. Healthcare utilisation was assessed to estimate the percentage of adults who engaged in AID who also had recently accessed healthcare for a routine check-up. This was measured by the question ‘About how long has it been since you last visited a doctor for a routine check-up? (A routine check-up is a general physical examination, not an examination for a specific injury, illness or condition.)’ Answers were recorded as being within the past 12 months, 2 years, 5 years or ≥5 years ago.
Analyses were carried out separately for each year. Results were weighted using the BRFSS-provided weights, cluster and stratification variables to make results nationally representative. National AID 30-day prevalence was estimated using the percentage of respondents who reported any AID in the previous 30 days. Annual estimates of AID episodes per respondent were calculated by multiplying the respondent’s reported episodes in the preceding 30 days by 12. For the 28 respondents (8 in 2014, 6 in 2016 and 14 in 2018) who reported more than one AID episode daily, annualised AID episodes were truncated at 360 (which is equivalent to 30 AID episodes per month). Annual rates of AID episodes and corresponding 95% CIs were then calculated by dividing the annual number of AID episodes by the respective weighted population estimate from BRFSS for the respective year (2014, 2016 or 2018). Each rate’s SE was used to calculate CIs and was approximated using Taylor series linearisation (also called the ‘delta method’).20 Annual AID episode rates were reported per 1000 population. National AID prevalence, number of episodes and rates per 1000 population were stratified by demographic and behavioural characteristics. Data analysis was completed using the complex sampling survey procedures in SAS V.9.4.
The analysis included over 1 million respondents from the 50 US states and District of Columbia who had non-missing AID information (426 910 in 2014, 448 062 in 2016 and 405 074 in 2018).
AID prevalence, number of episodes and rates
Nationally, 1.7%, 2.1% and 1.7% of adults in the years 2014, 2016 and 2018 reported having engaged in AID during the previous 30 days (tables 1–3).
On average, 57% of those who reported AID indicated one episode in the past 30 days, 24% indicated two episodes, 12% indicated 3–5 episodes and 7% reported that they had driven impaired ≥6 times over the past 30 days (data not shown). The estimated national annual number of AID episodes varied across years (2014: 111 million, 2016: 186 million, 2018: 147 million) and represented 3.7 million, 4.9 million and 4.0 million adults, respectively. The rate of AID episodes per 1000 population was highest in the year 2016 (rate=741, 95% CI 676 to 806) compared with 2014 (rate=452, 95% CI 412 to 492) and 2018 (rate=574, 95% CI 491 to 657).
AID by demographic and behavioural characteristics
In each year, AID was most common among men, people who binge drink and people who did not always use a seatbelt (tables 1–3). Men accounted for an overwhelming percentage of AID episodes (80% in 2014, 70% in 2016 and 80% in 2018; data not shown). Similarly, people who engaged in recent binge drinking accounted for 85%, 80% and 86% of all AID episodes in 2014, 2016 and 2018, respectively (data not shown). Those who reported more binge drinking reported more AID episodes. For example, in 2014, the 4% of adults who reported binge drinking at least four times per month accounted for 58% of AID episodes. This was true in 2016 and 2018 where 4% and 5% of those who reported binge drinking at least four times a month accounted for 55% and 65% of AID episodes in each respective year. People who reported not always wearing a seatbelt had an annual AID rate four times higher in 2014 and 2016 and six times higher in 2018 than those who always wore a seatbelt.
Reported AID varied by other characteristics as well. Regardless of gender and year, AID rates were highest among people aged 21–34 years and then decreased with age. Married adults, particularly married male adults, tended to have lower AID rates compared with those who were coupled, previously married or never married. There were no significant differences in AID rates by race/ethnicity, education level or household income no matter the year or gender. Among those engaging in AID, 60% reported seeing a doctor for a routine check-up within the past year (data not shown). Another 16% had a check-up between 1 and 2 years prior (data not shown). Among respondents who reported recent binge drinking, 62% reported a routine check-up within the past year (data not shown). Finally, among those reporting recent AID and recent binge drinking, 57% had a check-up within the past year (data not shown).
Discussion and public health implication
AID continues to be prevalent in the USA, and the majority of AID episodes during 2014–2018 occurred among men and those who engaged in recent binge drinking. AID prevalence and episode rates were also higher among those aged 21–34 years compared with older ages and among those who did not always wear seatbelts compared with those who always wear seatbelts.
These 2014, 2016 and 2018 BRFSS results are similar to previously published 2012 BRFSS results. In 2012, 2014, 2016 and 2018, 1.8%, 1.7%, 2.1% and 1.7% of adults engaged in AID. This translated to 4.2 million adults, 3.7 million adults, 4.9 million adults and 4.0 million adults engaging in 121 million annual AID episodes, 111 million episodes, 186 million episodes and 147 million episodes during each of the 4 years.18 Rates across the 4 years were 505, 452, 741 and 574 per 1000 population.18 Similar to 2014–2018, in 2012, men accounted for 80% of AID episodes and respondents who reported binge drinking accounted for 85% of episodes.18 Taken all together, there were slight differences in AID across these years with a peak in AID prevalence and number of episodes in 2016, but no clear trend across the years 2012, 2014, 2016 and 2018. This roughly correlates with national annual motor vehicle crash death data that suggest crash deaths and the percentage of them related to AID have remained relatively constant over the years 2012–2018.1 3–6 It is unclear what might be behind the peak in AID in 2016. Changes in AID can be influenced by changing economic and societal factors (like economic recessions). Preliminary data show an increase in AID-related crash deaths in 2020 (during the COVID-19 pandemic), which might signify an associated increase in 2020 BRFSS AID rates.21
AID-related deaths are preventable via proven strategies. To reduce AID, states and communities can consider implementing or scaling up effective interventions such as expanding the use of publicised sobriety check points; enforcing blood alcohol concentration (BAC) laws and minimum legal drinking age laws; requiring ignition interlocks for all persons convicted of AID and increasing alcohol taxes.22 Because a significant proportion of adults engaging in AID also does not always wear a seatbelt, primary seatbelt laws that cover all passengers might decrease AID-related crash mortality. Increasing seatbelt use among those engaging in AID is particularly important because alcohol not only increases the risk of a crash but also increases the risk of injury or death in a crash.23–25
Promising strategies that have shown effectiveness in other countries, when implemented, could decrease AID and subsequent crash deaths. The National Transportation Safety Board recommended lowering the BAC limit in the USA for drivers from 0.08 to 0.05 to reduce crashes, injuries and deaths caused by AID.26 A meta-analysis estimated that 1790 lives would be saved each year if all US states adopted a 0.05 BAC limit.27 Most high-income nations have already enacted a 0.05 illegal BAC limit, and these nations have lower motor vehicle crash fatality rates than the USA.28 Because our results showed that AID rates were highest among people aged 21–24 years (followed closely by people aged 25–34 years), future strategies that work among young adults are warranted. Although consuming alcohol is generally illegal in the USA for anyone under the age of 21 years, 1.1%, 1.5% and 1.5% of people aged 18–20 years reported engaging in AID during 2014, 2016 and 2018, suggesting the need to support strategies that prevent alcohol use and AID among young adults. It is unclear what effects ride share companies (eg, Uber and Lyft) might have on AID, and this topic deserves evaluation. Studies have shown mixed results with one showing that rideshare operations decreased alcohol-involved crashes only in certain cities29 while another showed no impact of rideshare services on alcohol-specific crash deaths.30
We found that three-quarters of people who engaged in AID attended a routine check-up with a doctor within the previous 2 years. This was also true for those who engaged in recent binge drinking and those who engaged in binge drinking and AID. Although not all people will accurately report their alcohol use, routine check-ups offer opportunities for healthcare providers to inquire about and discuss alcohol use and alcohol-related risky behaviours like AID. Alcohol screening and brief intervention (SBI), recommended by the US Preventive Services Task Force for all adults in primary care, is effective at identifying and reducing risky drinking behaviours in the primary care setting.31 Alcohol SBI guidelines recommend either of two brief screens.32 33 Healthcare staff can then initiate conversations on drinking limits and apply brief interventions34 tailored to individual patients’ motivations. The SBI intervention step is important but often overlooked. Although most people visiting their doctor are asked about alcohol consumption and binge drinking, most who report binge drinking receive no advice about how to reduce their drinking.35
The AID prevalence, episodes and rates reported here are likely underestimates of true AID prevalence in the USA for several reasons. First, BRFSS surveys only those aged ≥18 years, so AID episodes of younger drivers are not included. Second, BRFSS respondents were asked about times when they thought they had had too much to drink, and it is possible that respondents had times where they were impaired but did not recognise it. This might be particularly true for those with a history of AID.36 Third, respondents could have felt a social stigma associated with AID, which caused them to underreport AID. The 2018 National Survey on Drug Use and Health reported that 8% of the US population aged ≥16 years (which is an estimated 20.5 million people) reported driving under the influence of alcohol in 2018.37 This estimate is roughly five times greater than the 2018 BRFSS estimate. This is likely partly because the National Survey on Drug Use and Health included 16 and 17-year-old participants and partly because it used Audio Computer-Assisted Self-Interview software (ie, computer-administered survey) methodology, which might heighten respondents’ sense of privacy and, thereby, increase their willingness to report AID compared with BRFSS’s telephone survey methodology.38 39 Another study similarly found that passengers who report riding with a drinking driver might provide a more accurate prevalence of AID than drivers.40 Although BRFSS estimates are likely underestimates, they can help describe the magnitude of AID in the USA. Additionally, other characteristics that BRFSS collects can help describe those who report AID to facilitate prevention efforts.
There are other limitations to this analysis. First, we assumed that what people reported over the past 30 days represented their experience over the past 12 months. This might not be a reasonable assumption, especially because AID is more common during certain seasons and holidays. However, BRFSS interviews took place year-round, likely minimising any seasonal bias. Second, BRFSS only asked about the number of times a person drove after consuming too much alcohol and not the total miles travelled or length of trip time, which might be more relevant but less precise (because it might be harder for people to self-report accurately) measures of exposure. Third, the BRFSS AID question asked whether respondents perceived that they had had too much to drink before driving, and it is unclear how this might relate to crash risk or blood alcohol concentrations. In the USA, it is illegal for a driver to have a blood alcohol concentration of 0.08 g/dL or higher, except in Utah where it is illegal to have a blood alcohol concentration of 0.05 g/dL or higher. However, studies have shown that even small amounts of alcohol (eg, <0.08 g/dL) can reduce motor skills and reaction time.22 41 Finally, there could be unknown differences between people who report AID and people who die or are injured in an AID-related crash.
AID during the years 2014, 2016 and 2018 was prevalent and linked to other risky behaviours including binge drinking and not always wearing seatbelts. AID is preventable. Because 29% of motor vehicle deaths in 2018 involved an alcohol-impaired driver, eliminating or reducing AID could potentially reduce crash-related deaths by 20%–30%, saving roughly 7000 to 11 000 lives each year.1 In addition to saving lives, the impact would also be felt by reduced injuries and burdens on healthcare and emergency response systems. States and communities can consider enacting and enforcing AID-reducing strategies at a population-level while healthcare providers in primary care settings can consider addressing AID at an individual level.
What is already known on the subject?
Alcohol-impaired driving is a risk factor for traffic crashes and their resulting injuries and deaths.
In 2012, an estimated 1.8% of adults (or 4.2 million adults) in the USA reported alcohol-impaired driving within the past 30 days
What this study adds
More recent estimates from the years 2014–2018 indicate that reported alcohol-impaired driving remains prevalent. An estimated 1.7%, 2.1% and 1.7% of adults (or 3.7 million, 4.9 million and 4.0 million adults) in the USA reported alcohol-impaired driving in 2014, 2016 and 2018.
Alcohol-impaired driving was more common among men and among people who binge drink.
Data availability statement
Data are available in a public, open access repository. Not applicable.
Patient consent for publication
Contributors Author EKSS conceived of and designed the study. Authors VB and ACS performed the statistical analyses. Author VB wrote the manuscript. Authors VB, ACS, and EKSS critically revised the manuscript. VB is the guarantor for this work.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Disclaimer The findings and conclusions in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.