National Academies Press: OpenBook

Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem (2018)

Chapter: Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver

« Previous: 8 Conclusion
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

A

Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses
1

Charles DiMaggio, Ph.D., M.P.H.,
Katherine Wheeler-Martin, M.P.H., and Jamie Oliver

SUMMARY2

The consequences of alcohol-impaired driving continue to affect the United States. A review of the current literature and analyses of recent data indicate a need for renewed surveillance across the spectrum of potential interventions, including law enforcement, engineering and technology, education and behavioral change, built environment, enactment and evaluation of policies, and emergency trauma care. Among these so-called Es of motor vehicle crash injury prevention, alcohol itself (ethanol) remains critically important. There has been considerable success in addressing the role alcohol plays in motor vehicle crash injury risk in the United States, but alcohol has been persistently present in nearly a quarter of fatal crashes for the past two decades. Initial decreases of

___________________

1 This background paper was commissioned at the direction of the Committee on Accelerating Progress to Reduce Alcohol-Impaired Driving Fatalities to offer insight into the data and surveillance systems that could inform alcohol-impaired driving interventions in the United States. In addition, this paper provides the descriptive epidemiology of fatal and nonfatal alcohol-involved crashes in the United States over time, describes trends in patterns of alcohol consumption and alcohol-impaired driving over time, and identifies the strengths and weaknesses of existing data sources that inform this epidemiology.

2 This is an abridged version of the paper. A full version including all tables and figures can be found in the public access file for the Committee on Accelerating Progress to Reduce Alcohol-Impaired Driving Fatalities. Available by request from the National Academies of Sciences, Engineering, and Medicine’s Public Access Records Office (PARO@nas.edu).

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

30 to 40 percent dating from the 1990s have slowed to 5 to 10 percent or leveled off entirely.

Law enforcement has been, and remains, a cornerstone of alcohol-impaired driving control and prevention, but enforcement activity varies widely across states, and there is evidence of an overall decline in alcohol-impaired arrest rates across the nation. While alcohol-impaired driving accounts for over 10 percent of all arrests reported to the Federal Bureau of Investigation’s (FBI’s) National Incident-Based Reporting System (NIBRS), arrest rates and enforcement activity vary as much as 32-fold across states. The population-based rate of alcohol-impaired driving arrests reported to the NIBRS in 2014 decreased by 13 percent from a high in 2003.

Educational and behavioral interventions have been a foundational feature of the public health approach to alcohol-impaired driving since the earliest days of community-level activism that led to the formation of such groups as Mothers Against Drunk Driving. Identifying what behaviors are associated with which groups is essential to this effort. Some features of alcohol-impaired driving behavior remain unchanged. Men are still three to four times as likely to report recent incidents of alcohol-impaired driving, and the highest population-based rates continue to be among younger drivers. But there have been notable recent changes in behaviors among some groups, indicating an evolution in risks posed by alcohol-impaired driving.

Teens and young adults are driving less, which is reflected in decreases in alcohol-impaired driving arrests, crashes, and fatalities. However, this group may now be at increasing risk of injury as passengers, and behaviors leading to decisions to ride as a passenger with alcohol-impaired drivers deserve increasing attention. The rate of women in the United States reporting a recent arrest for alcohol-impaired driving increased 40 percent between 2002 and 2014. Statistics reported to the FBI over the same time period show a similar increase. This increase may reflect, in part, an increase in marketing of alcohol to women. Motorcyclists, particularly those 55 to 64 years old, have experienced a notable increase in alcohol-related crash fatalities in the United States. In Florida, between 2011 and 2013, the fatality rate for motorcyclist crashes in which alcohol was not involved was 5 percent; when alcohol was involved, the fatality rate was 21 percent. In Nebraska, between 2002 and 2013, all crashes with a driver suspected of using alcohol or with a positive blood alcohol concentration (BAC) report decreased 25 percent, but rates among motorcyclists remained stubbornly persistent. Surveys consistently indicate that alcohol-impaired driving is associated with other risky behaviors such as binge drinking and not using seat belts. Investigating and

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

addressing these interactions may help better tailor interventions that result in decreased morbidity and mortality.

Much of what we know about alcohol-impaired driving behavior comes from national surveys, such as the Centers for Disease Control and Prevention’s (CDC’s) Behavioral Risk Factor Surveillance System (BRFSS), which is among the most well-accepted and reliable sources of data on health-related behavior in the United States. Surveys can be enhanced to include questions, on alcohol-impaired driving on an annual basis. Additional survey questions. such as whether a person has been stopped at a sobriety checkpoint or been assigned to use an ignition interlock system, can provide information that may help track interventions. But there is a need for objective data to balance survey results. The FBI’s NIBRS has the potential to provide important information on alcohol-impaired driving behavior and enforcement in the United States. It is limited in that it is not yet fully implemented across all states or representative of the entire United States. Roadside surveys conducted by the National Highway Traffic Safety Administration (NHTSA) provide perhaps the most valid and reliable estimates of alcohol-impaired driving behavior in the United States, but they are only conducted sporadically.

There is a need for additional, updated research addressing environmental factors associated with alcohol-impaired driving. Approximately 60 percent of alcohol-related motor vehicle crash deaths in the United States occur in rural areas. Adjusting for vehicle miles traveled indicates a risk three times greater than that of urban areas. Possible explanations for the rural–urban difference may include differences in access to alcohol on a per capita basis, and limited access to low-cost public transportation in rural areas. The availability of large data sources, which can be linked, and advances in small-area analyses can allow for better characterization of that risk and targeting of interventions. The NHTSA Fatality Analysis Reporting System (FARS), the premier and most successful source of alcohol-impaired driving fatality statistics in the United States, can contribute to that effort by consistently including geolocations. Most states participating in the NHTSA’s State Data System (SDS) already collect and provide information about alcohol involvement in motor vehicle crashes. However, there is substantial variation in how this information has been collected over time and from state to state. Beginning in 2006, NHTSA began standardizing data elements across several of its data systems. Increased state participation in that system using consistent definitions will contribute to a better understanding of the actual incidence of alcohol-related motor vehicle crash fatalities in the United States.

There have been few studies looking explicitly at ethnic and racial patterns of alcohol-impaired driving behavior. More information is needed to better inform and target control efforts. Similarly, studies of the

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

effectiveness of engineering and technological interventions for alcohol-impaired driving have been largely limited to ignition interlock systems. Despite the tremendous potential of self-driving and assisted driving technologies to affect alcohol-impaired driving outcomes, the research literature on the topic is virtually nonexistent. Lastly, while most analyses of alcohol-impaired driving in the United States rightly address the overall public health impact through absolute numbers of fatalities, additional analyses attempting to capture risk, as measured through rates of both crashes and fatalities, can better inform prevention and control through the use of estimates of vehicle miles driven to serve as denominators for analyses.

In summary, much has been achieved in controlling and preventing the consequences of alcohol-impaired driving in the United States, but much remains to be done. Recent evidence and analyses can help point the way to opportunities for additional advancements. Further success will require a renewed commitment to the prevention and control of alcohol-impaired driving in the United States with resources targeted at those individuals and areas most at risk.

ABOUT DATA SOURCES

This paper reviews administrative secondary data sources with information relevant to alcohol-impaired driving. CDC’s BRFSS is among the most well-accepted and reliable sources of health-related behavior in the United States, and it should be enhanced to include a question on alcohol-impaired driving on an annual basis. Summary information about key data sources is listed below.

  • Behavioral data characterized by self-reports may be biased.
  • The NIBRS has the potential to provide important information on alcohol-impaired driving behavior and enforcement in the United States. It is limited in that it is not yet fully implemented by all states nor representative of the entire United States.
  • Most states participating in the NHTSA’s SDS collect and provide information about alcohol involvement in motor vehicle crashes. However, there is substantial variation in how this information has been collected over time and from state to state. Consistent definitions and greater participation among states can provide valuable surveillance information.
  • The NHTSA’s FARS is the premier source of alcohol-impaired driving fatality statistics in the United States. Data have been collected in the FARS from all 50 states since 1975.
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
  • Data on alcohol involvement in the FARS are often missing, resulting in likely undercounts of the actual incidence of alcohol-related motor vehicle crash fatalities.
  • The U.S. Department of Transportation’s National Household Travel Survey provides reasonable estimates of vehicle miles driven to serve as denominators for analyses.

Behavioral Data

The Behavioral Risk Factor Surveillance System

The BRFSS is an annual nationwide survey conducted by CDC and is perhaps the most important, reliable, and long-standing source of information on health behavior in the United States. Since 2006, every other year, the BRFSS has included an explicit question on alcohol-impaired driving that asks, “During the past 30 days, how many times have you driven when you’ve had perhaps too much to drink?” Respondents answer with the number of times, from 0 to 76. There is an annual question on the number of alcoholic beverages a person consumes each day, which is used to calculate the total number of alcoholic beverages consumed in a month, and determine whether an individual can be considered a binge or heavy drinker. There is an annual question on seatbelt use (“How often do you use seat belts when you drive or ride in a car?”), which can be evaluated in relation to alcohol use.

Among the strengths of the BRFSS is the well-accepted reliability and validity of the results over a long period of time. Among the weaknesses are (1) because it is a phone-based survey, some of that strength is challenged by changes in phone use patterns, (2) self-report of what might be considered deviant behavior is likely to result in bias, (3) questions pertaining to alcohol-impaired driving are only present every other year and therefore provide data for 6 years, making it difficult to establish trends, and (4) results are only valid at the national and state levels, making local behavior patterns difficult to assess. The usefulness of the BRFSS to track alcohol-impaired driving in the United States can be enhanced by including a related question every year. Potential biases in self-reports can be at least partially addressed through the inclusion of such methods as randomized response or item count techniques (Miller et al., 1986; Warner, 1965).

Similar to the BRFSS is the Youth Risk Behavior Surveillance System (YRBSS), which monitors health behaviors among U.S. youth and young adults. The survey is conducted every other year and dates to 1991. Separate versions of the survey are administered to high school and middle school students. Among the six specific health behaviors monitored are

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

alcohol and other drug use, and there are two specific questions relating to alcohol-impaired driving:

  1. Have you ever ridden in a car driven by someone who had been drinking alcohol?
  2. During the past 30 days, how many times did you ride in a car or other vehicle driven by someone who had been drinking alcohol?

Among the strengths of the YRBSS are that it has consistently included questions on alcohol-impaired driving that allow for reliable estimates of trends and addresses risk among a vulnerable population. Weaknesses of the YRBSS include (1) data may not be available for all states, (2) limited geographic specificity similar to the BRFSS, and (3) biases associated with self-reported behavior likely similar to the BRFSS, although there may be some added level of anonymity in the way the YRBSS is administered.

National Survey on Drug Use and Health

The National Survey on Drug Use and Health is a yearly survey of “national and state-level estimates on the use of tobacco products, alcohol, illicit drugs (including nonmedical use of prescription drugs) and mental health in the United States” conducted by the Substance Abuse and Mental Health Services Administration. Included among the questions is “In the past 12 months, were you arrested and booked for driving under the influence of alcohol or drugs?” A strength of this data source is that it is a large survey of approximately 70,000 individuals conducted yearly since 1971, making it an effective resource to validly track trends at the state and national levels. Unlike the BRFSS, the survey includes individuals in institutional settings such as shelters, rooming houses, and dormitories. Like the BRFSS, the survey is subject to bias in that it relies on self-reported data and is not powered to assess outcomes at smaller geographic levels.

National Incident-Based Reporting System

The FBI maintains the NIBRS as a means of quantifying criminal activity in the United States. In contrast to the FBI’s longer-standing traditional Uniform Crime Reporting (UCR) statistics, the NIBRS collects data on a wider array of crimes and goes into greater detail on the circumstances surrounding offenses. Unlike the UCR, and among its weaknesses, the NIBRS is not yet nationally representative. As of June 2012, 32 states were participating. According to the U.S. Department of Justice, “In 2015, 6,648 law enforcement agencies, representing coverage of more than 96

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

million U.S. inhabitants, submitted NIBRS data. While not yet nationally representative (there are no estimates for agencies that did not submit NIBRS data), this coverage represents 36.1 percent of all law enforcement agencies that participate in the UCR program” (FBI, 2016). The target date for full national representation is 2021.

Among the strengths of these data are that they are (for the states represented) essentially a census of crimes, rather than a survey, and are less sensitive to statistical heterogeneity. Driving while impaired (DWI) has been a reportable offense since the inception of the NIBRS. It is classified as a so-called Group B offense for which only arrestee data are reported. Among the weaknesses of these data as a source of surveillance information on alcohol-impaired driving is that they are as much a reflection of enforcement activity and availability of resources as they are an estimate of actual behavior.

National Roadside Survey of Alcohol and Drugged Driving

NHTSA has conducted the National Roadside Survey of Alcohol and Drugged Driving five times since 1973. The most recent iteration was conducted in 2013 and 2014 and consisted of a sample size of approximately 9,000 persons in 60 U.S. cities. The major strength of this survey is that results are based on breath, saliva, and blood testing rather than self-reporting, increasing the validity of the conclusions. Among the weaknesses is the intermittent nature of the survey, making it difficult to assess trends.

Crash Data

The State Data System

The primary source of alcohol-related motor vehicle crash injury data is the NHTSA SDS. The SDS is based on data from police crash reports submitted by participating states. Police crash reports are completed by police officers at the scene of motor vehicle crashes and contain information about the crash, the vehicles involved, and the motorists and nonmotorists (e.g., pedestrians, cyclists) involved in the crash. Each state determines its own reporting criteria. Generally, crashes are reportable if they involve injuries or deaths. Other common reporting criteria include damages in excess of a designated cost or damages requiring a vehicle to be towed away from the scene. While the specific nature and quantity of information collected at the crash site varies from state to state (and over time within each state), the SDS applies a common nomenclature to

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

field names when processing and storing state datasets to aid in research efforts across states.

As of June 15, 2017, there were 34 states participating in the SDS. Access to state data requires special permission from each individual state data owner. Not all states allow NHTSA to rerelease their data.

Most states participating in the SDS collect and provide information about alcohol involvement in motor vehicle crashes. However, there is substantial variation in how this information has been collected over time and from state to state. The range of language and specific indicators used to identify alcohol use include

  • Individual had been drinking.
  • Individual was/appeared to be under the influence of alcohol.
  • Individual was/appeared to be intoxicated.
  • Individual received an alcohol test.
  • Individual BAC percentage ranging from 0.00% (0%) to 0.99% (9%).
  • Alcohol was a contributing factor, circumstance, or cause of the crash.
  • Alcohol was present in the vehicle/at the scene.
  • Individual was issued a violation for alcohol use, being under the influence, etc.

Alcohol information is almost always available for drivers. Very few states collect or report alcohol information for passengers. Some states provide indicators at the crash level that may be used to readily identify crashes that involved alcohol on some level. Alcohol indicators are more commonly reported at the individual level, which can allow for analysis by role (driver, pedestrian/cyclist), age, and gender.

The strengths and weaknesses of alcohol-related motor vehicle crash data in the SDS vary considerably by state. For example, before 2011, Florida SDS datasets included alcohol information for individuals using the language “under the influence.” Beginning in 2011, Florida began including information about alcohol testing and test results, and added a flag for “alcohol use suspected,” which captures a wider range of crashes with alcohol involvement. By contrast, Nebraska data are based on “officer deemed alcohol related to crash,” with additional fields available for whether an alcohol test was performed and whether results are known.

By contrast, New York State SDS datasets identify alcohol information only in the context of violations issued, which essentially pertain to drivers only. Violation data files were added to New York’s SDS files beginning as early as 2002. There are a number of different types of violations

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

with language ranging from alcohol use, impairment, intoxication, BAC above 0.08% (CDL) or 0.1% (private vehicle), and open container of alcohol.

Because SDS data include injuries as well as fatalities, an additional strength of this data source is that fatality rates can be calculated and compared across crash characteristics. Additionally, the number of participating states has grown over time and now covers the majority of the continental United States. Most states provide a substantial array of variables related to crash circumstances, vehicles involved, and some information about persons involved.

While the SDS provides a common naming structure and language, as noted, a limitation is that there are many differences in how variables are defined and how information is collected and reported from state to state. Also, states have also changed their coding and reporting practices over time, requiring careful attention to the coding manuals. In particular, as noted above, alcohol involvement is recorded in numerous ways and makes state-to-state comparisons problematic. Finally, geographical information in the SDS is limited to city and county.

National Automotive Sampling System General Estimates System

NHTSA’s National Automotive Sampling System General Estimates System (NASS GES) is based on a sample of police crash reports. The data are used to track trends and identify emerging issues. Starting in 2006, NHTSA began standardizing data elements across the FARS and the NASS GES, increasing the usefulness of the system.

Fatality Data

Fatality Analysis Reporting System

The primary data source available in the United States for traffic-related mortality is the NHTSA’s FARS. When combined with U.S. Census population estimates, or person-miles traveled estimates such as those available from the National Household Travel Survey, and vehicle miles traveled estimates from the Federal Highway Administration, reliable motor vehicle mortality-related rates can be calculated.

The FARS is a national census of fatal motor vehicle crashes (MVCs) and is maintained by the National Center for Statistics and Analysis, a component of NHTSA. Initially established in 1975, the FARS contains detailed information on all crashes that occur on U.S. public roadways that result in the death of one or more motorists or nonmotorists within 30 days. FARS data are acquired from police reports, state administrative files, and medical records from all 50 states and are analyzed by data

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

analysts using standardized protocols. As a result, FARS data elements generally have relatively little missing data and are well documented and maintained.

For each fatal crash, the FARS provides detailed information related to the crash circumstances as well as the people and vehicles involved in that crash. The FARS lists the county, city, state, and longitude and latitude of each crash and also includes roadway information (road surface type, number of lanes, speed limit, traffic control device), and light and atmospheric conditions. The time of the crash and emergency medical services transportation information (arrival at the scene and to the hospital) are also included. In addition, driving maneuvers avoiding and contributing to each crash are documented. Data on each vehicle involved include the make, model, and model year as well as less detailed variables to aggregate and compare crashes by vehicle type.

Data on all motorists and nonmotorists involved in each crash, including age, sex, seating position, vehicle occupant restraint use, and injury severity, are provided. For drivers, information on prior crashes, license suspensions, and other driving violations within 3–5 years preceding the crash is available. Of note, certain information such as race and ethnicity, which is collected from death certificates, is only available for fatally injured individuals.

The FARS also has data on alcohol and drug intoxication that are gathered from breath, blood, and urine tests, as well as police behavioral assessments. Since these data are based on the extent of testing performed by law enforcement, alcohol involvement data are often missing. As a result, the FARS may undercount the actual incidence of alcohol-related motor vehicle crashes, and testing rates are a potential confounding variable for regional comparisons of alcohol-related crashes.

Strengths of the FARS The various strong points of the FARS are listed below.

  • Data have been collected in the FARS from all 50 states since 1975, allowing for long-term trend analysis.
  • Significant data are available to analyze crashes at the crash level, vehicle level, and person level, and these data are quite detailed at every level.
  • For drivers, detailed information is available on a driver’s previous traffic and driving violations prior to the crash.
  • Data elements are well documented and maintained.
  • The majority of FARS variables have very little missing data.
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

Weaknesses of the FARS The various weak points of the FARS are listed below.

  • The FARS only has information on crashes resulting in at least one fatality within 30 days of the crash.
  • Data on alcohol involvement are often missing, meaning the FARS may undercount the actual incidence of alcohol-related motor vehicle crashes.
  • Police testing for illicit substances may vary between different states and regions and is a confounding variable for state-level comparisons.
  • Data on race and ethnicity are collected from death certificates, meaning that it is only available for individuals fatally injured in crashes. Even among deceased individuals in alcohol-related crashes, race and ethnicity information is missing for 14 percent of persons.
  • The FARS occasionally has errors in data element coding, but these errors are well documented and can be worked around.

Multiple imputation of alcohol data in FARS Missing data are an issue in most rigorous data analyses and are particularly problematic in alcohol-related crash analyses. While it is the standard for such analyses, more than half of FARS records are missing BAC data (Subramanian and Utter, 2003). Over the years, NHTSA has taken several approaches to estimating missing BAC values in the FARS. Prior to 2001, a linear discriminant method was used to estimate the probability that a driver or nondriver (pedestrian or cyclist) had a BAC of 0.0, 0.01−0.09, or >0.10 g/dl. Starting in 2001, NHTSA moved to multiple imputation methods (Subramanian, 2002). The algorithm generates 10 estimates for BAC values for each missing BAC observation. The models used to estimate missing values include variables such as age, sex, injury severity, and day and time of crash that are likely to be associated with BAC. The final imputed value is a combination of the 10 estimates and includes a measure of the uncertainty, which is added to variances for overall point prevalence. Multiple imputation methods are superior to some other missing data methods, and the methods pioneered by NHTSA have been applied to missing drug testing data in truck crashes (Brady et al., 2009). However, they cannot substitute for actual testing results. Data imputation is least biased when values are missing at random, which is unlikely in the case of alcohol-related fatalities. It is important to evaluate results based on imputed values in comparison to complete-case analyses and explain any discrepancies (Sterne et al., 2009).

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

Clinical Data

Nationwide Emergency Department Sample

The Agency for Healthcare Research and Quality (AHRQ) Healthcare Cost and Utilization Project (HCUP) maintains the Nationwide Emergency Department Sample (NEDS) (AHRQ, 2015). The HCUP is a group of inpatient and outpatient files created by AHRQ. NEDS is the largest single publicly available emergency department (ED) database in the United States. It is based on a 20 percent stratified single-cluster sample of all U.S. hospital-based EDs. As of 2012, 30 states participated in NEDS, accounting for 66 percent of all national ED visits. Each year’s core file contains 100 percent of visits from sampled hospitals. The most recent NEDS database contains about 30 million ED records. Hospitals are defined as nonfederal general and specialty hospitals, including public hospitals and academic medical centers. Additional stratification variables include geographical area, urban/rural, ownership, trauma center and teaching status, and bed size.

Motor vehicle crashes can be identified using external cause of injury e-codes. Alcohol intoxication can be determined using International Classification of Diseases (ICD)-9 diagnostic codes. Combining the two variables can allow for the identification of alcohol-related motor vehicle crashes. Among the strengths of the NEDS is the large, comprehensively representative sample with few missing data that can allow for the analysis of rare outcomes. Among the weaknesses is the difficulty in working with the large datasets required for multiyear analyses. An important bias common to clinical datasets based on administrative diagnostic codes is that diagnoses of alcohol intoxication are almost invariably limited to the most intoxicated individuals, resulting in important selection bias.

Web-Based Injury Statistics Query and Reporting System

CDC’s Web-based Injury Statistics Query and Reporting System (WISQARS) is an online, publicly available resource for both fatal and nonfatal injury-related statistics. Fatality data are drawn from CDC’s National Center for Health Statistics, National Vital Statistics System Mortality Data, which are themselves based on death certificate ICD-10 assignments. The ICD-10 coding system allows for inclusion of alcohol as an attributable cause (Codes X45, Y15, T51.0, T51.1, T51.9) of motor vehicle crashes (Codes E820–E825), but this combination is not routinely reported as part of the WISQARS. Injury data are drawn from the National Electronic Injury Surveillance System All Injury Program (NEISS-AIP) operated by the Consumer Protection Safety Commission. The NEISS is a national probability sample of approximately 100 hospitals in the United

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

States and its territories drawn from a commercially purchased sampling frame (Schroeder and Ault, 2001).

Data Linkage: Crash Outcome Data Evaluation System

NHTSA developed the Crash Outcome Data Evaluation System (CODES) to help researchers link crash records with clinical records. The CODES uses a match weight test statistic and has been used in a number of motor vehicle crash injury studies (see, for example, Clark et al., 2004; Loo and Tsui, 2007; Lopez et al., 2000). Advances in probabilistic and fuzzy matching methods offer additional approaches to matching datasets (Lujic et al., 2008).

The CODES was created by NHTSA in the early 1990s to inform motor vehicle safety legislation. NHTSA initially funded 16 states to develop and implement the probabilistic data linkage algorithms for motor vehicle crash data sources. Some states expanded their efforts to link crash data with hospital discharge records, emergency medical transport call reports, and trauma registry records. Many of these efforts rely on deterministic linkage using identifiers not generally available to nongovernment researchers. Commercial software that replicates the CODES methodology is also available.

Denominator Data

Among the more challenging aspects of measuring alcohol-related motor vehicle crashes in the United States is determining an appropriate denominator to use to calculate comparable rates across time and place. While absolute numbers of fatalities, crashes, and incidents provide important information on the effect of alcohol-impaired driving in the United States, additional attention to normalizing such numerators by appropriate denominators can help better define risk at a time when resources need to be most efficiently deployed. A common approach is to use census data for population denominators, but other approaches based on travel patterns are possible.

National Household Travel Survey

The U.S. Department of Transportation conducts periodic national surveys about how people travel across the United States in their daily lives; the most relevant to alcohol-impaired driving is the National Household Travel Survey (NHTS). Data are available online for the years 1995, 2001, and 2009. The 2016 survey was recently administered and should be available in the near future. Data collected include

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
  • Purpose of the trip (work, shopping, etc.);
  • Means of transportation used (car, bus, subway, walk, etc.);
  • How long the trip took (i.e., travel time);
  • Time of day when the trip took place;
  • Day of the week when the trip took place;
  • If a private vehicle trip, number of people in the vehicle (i.e., vehicle occupancy);
  • Driver characteristics (age, sex, worker status, education level, etc.); and
  • Vehicle attributes (make, model, model year, amount of miles driven in 1 year).

The 1995 and 2001 surveys provide national estimates, and the 2009 survey provides both national- and state-level estimates, as it included a larger sample size. A full description of the NHTS methods and data use considerations may be found online.

To account for intervening years one can apply linear regression to NHTS point estimates from 1995, 2001, and 2009 to create annual estimates of drivers, vehicle miles driven, and person-miles traveled (total and subdivided miles traveled in privately owned vehicles, walking, and biking). Each estimate was stratified by age and sex and by race and ethnicity, and linear interpolation was conducted separately for each stratum.

Among the strengths of the NHTS is that it allows analysis of driver counts, miles driven, and person-miles traveled across a range of household and person characteristics as well as vehicle and trip characteristics, and that it is a large sample—a nationally representative survey.

Among its limitations are that race and ethnicity are applied to driver and household persons based on the primary household respondent. In one stratum, the linear regression produced a negative value for the first observation. In this case (bicycling miles traveled by men ages 75 and older), the intercept was manually adjusted such that the first value was the actual 1995 estimated value.

While 2009 survey estimates are valid at the state level and may be further stratified by demographic characteristics within state, some strata may contain small numbers and aggregation may be required. Because state-level data are not available before 2009, linear interpolation was only conducted at the national level to produce annual estimates.

Federal Highway Administration Highway Vehicle Miles Traveled

The Federal Highway Administration (FHWA) highway vehicle miles traveled (HVMT) provides annual summary statistics at the national and state level on registered vehicles, drivers’ licenses, highway miles

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

traveled, and highway user taxation. FHWA compiles the data as submitted by individual states. The data have been collected annually since 1945.

This paper uses annual highway miles traveled by state for SDS analyses at the state level. Urban and rural highway miles traveled were also used with FARS data on urban and rural fatal alcohol-related crashes.

Among the strengths of FHWA highway statistics is that the survey provides a nationally consistent measure of vehicle miles traveled. Among its limitations are that data are based on vehicle counts and therefore demographics of miles traveled are unavailable. A full description of FHWA highway statistics data use considerations may be found online.

Issues and Limitations of Denominator Data

While vehicle miles traveled (VMT) provides a valid approach to estimating risk based on a direct measurement of exposure, there are some methodological challenges in its use (see Table A-1). The NHTS and the FHWA measures are structured differently and cannot be easily or reliably combined. The NHTS, for example, may allow for age stratification, while the FHWA estimates can be broken down by urban or rural geography. More granular stratifications, such as combining age groups and urban versus rural status, are not a trivial problem and require careful use of one dataset or the other. Measuring change over time will generally require interpolation over a limited number of measurements (in some cases just two), with an underlying assumption of linearity that may not be supported.

TABLE A-1 Strengths and Limitations of Denominator Data

Data Type Strengths Potential Weakness/Bias
1. Count data Commonly reported
Estimates absolute public health impact
Allows for stratification
Cannot estimate risk
2. Census population Easy to obtain
Estimate risk
Age stratification
Geographic stratification
Estimates over time
Race/ethnicity data
Economic data
May not accurately represent population at risk
3. Vehicle miles traveled Accurately represents exposure Limited stratification variables
Limited local geography
Limited years of observation
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

ANALYSES

BRFSS Data Analysis

  • The number of adults in the United States reporting an instance of alcohol-impaired driving over the previous 30 days decreased about 38 percent from 2006 to 2014. There was an approximate 10 percent decrease between 2010 and 2014.
  • Men were three times more likely to report recent alcohol-impaired driving.
  • The youngest age groups were more likely to experience a decline in recent alcohol-impaired driving.
  • Much of the decrease in reported rates of alcohol-impaired driving in the past 30 days for all age groups (≥18 years) occurred between 2006 and 2008.

The mean number of times an adult (≥18 years of age) reported alcohol-impaired driving over the past 30 days for all the data years and all states was 0.09 (standard error [s.e.] = 0.02, 95% confidence interval [CI] 0.09, 0.10). This represented 189.9 persons (≥18 years old), per 10,000 surveyed U.S. adult population members (s.e. = 2.0, 95% CI 185.6, 194.1), reporting that they drove while alcohol impaired during the previous 30 days. This population-based count decreased 38.4 percent (s.e. = 0.07) from 251.4 per 10,000 in 2006 (95% CI 238.3, 264.4) to 154.9 per 10,000 in 2014 (95% CI 147.5, 162.3). There was a 10.5 percent (s.e. = 0.5) decline from 2010 to 2014.

The overall rate of respondents reporting at least one incident of DWI in the previous month varied by U.S. state and territory, with the highest rate of 414.1 per 10,000 (95% CI 334.6, 493.6) in Guam and the lowest of 62.7 per 10,000 in Utah (95% CI 51.9, 73.6) (see Figure A-1). There was some graphical indication of clustering of higher rates in the northern United States (see Figure A-2). There was some variation, but the general geographic patterns prevailed over time.

More than three times more male (293.9 per 10,000; 95% CI 286.0, 301.7) than female (91.5 per 10,000; 95% CI 87.8, 95.2) respondents reported at least one incident of DWI in the previous 30 days. There was a decline during the study period in the number of both male and female respondents reporting at least one incident of DWI. The rate per 10,000 of respondents reporting at least one incident of DWI in the previous 30 days was inversely related to age group. All age groups experienced a decline from 2006 to 2014 in the rate per 10,000 of respondents reporting at least one incident of DWI in the previous 30 days, with the steepest declines among the youngest age groups (see Figure A-3).

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
images
FIGURE A-1 Number of respondents reporting at least one instance of alcohol-impaired driving in the past 30 days per 10,000 target adult survey population by state or territory.
SOURCE: BRFSS data, 2006–2014.

National Survey on Drug Use and Health Data Analysis

  • There was a 27 percent decrease from 2002 to 2014 in the rate of persons reporting an arrest for alcohol-impaired driving in the past month, with a more modest and more variable decline of approximately 10 percent since 2011.
  • Men were four times more likely to report a recent arrest for alcohol-impaired driving.
  • The rate of women reporting a recent arrest for alcohol-impaired driving increased 40 percent between 2002 and 2014.
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
images
FIGURE A-2 Choropleth number of respondents reporting alcohol-impaired driving in the past 30 days per 10,000 target adult survey population by state.
SOURCE: BRFSS data, 2006–2014.
images
FIGURE A-3 Number of respondents reporting alcohol-impaired driving in the past 30 days per 10,000 target adult survey population by age group over time.
SOURCE: BRFSS data, 2006–2014.
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

There were 723,283 entries in the National Survey on Drug Use and Health for the years 2002 to 2014, representing a survey-adjusted total of 3,250,847,536 (s.e. = 9,889,020) observations. The yearly survey count mirrored the total U.S. population for the years 2002 to 2014.

The overall rate of persons responding yes to questions regarding whether they had been arrested for DWI in the preceding 12 months was 57.8 per 10,000 survey-adjusted target population (s.e. = 0.0001). The rate per 10,000 of responses of yes to questions regarding arrest for DWI showed a fairly steady pattern from 2002 to 2010, after which there was evidence of a decline. The rate declined 26.7 percent (s.e. = 8.9) from 65.9 per 10,000 in 2002 to 47.9 per 10,000 in 2014. The decline between 2011 and 2014 was more modest (9.4 percent) and more subject to chance (s.e. = 12.4). The rates for 19- to 29-year-olds were nearly three times those for other age groups, although there was graphical evidence of a decrease among 19- to 21-year-olds.

The rate for men answering yes to questions regarding recent DWI arrests was nearly four times that of women, 94.0 per 10,000 target population men (95% CI 88.7, 99.3) versus 23.9 per 10,000 women (95% CI 21.8, 26.0). There was graphical evidence that the rates for men responding yes to questions regarding DWI declined over the study period, while rates among women appeared to be increasing. The rate for male respondents declined 38.6 percent (s.e. = 10.6) from 118.2 per 10,000 in 2002 to 73.1 per 10,000 in 2014, while the rate for female respondents increased 43.7 percent (s.e. = 28.9), from 17.1 per 10,000 target population in 2002 to 24.4 per 10,000 in 2014.

The age-gender category with the highest overall rate of responding yes to questions regarding DWI in the previous year was men aged 18–25, which at 213.8 per 10,000 target population (95% CI 200.3, 227.2) was over three times that of the next highest category. It was also the age category with the steepest decline in rates over the study period, decreasing 43.3 percent (s.e. = 11.6) from 296.1 per 10,000 in 2002 to 169.6 per 10,000 in 2014.

Rates of positive responses to questions regarding DWI were higher in less population-dense core-based statistical areas (CBSAs). These differences were statistically significant (p = 0.0001) in a survey-adjusted analysis of variance comparing CBSAs for a positive response to DWI questions. CBSAs of greater and less than 1 million people both experienced declines between 2002 and 2014 in the rates of individuals responding yes to questions about DWI. Although non-CBSA geographic segments had a 38.8 percent (s.e. = 44.2) increase in the rate of positive responses to questions regarding DWI from 2002 to 2014, there was considerable variability in the yearly data.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

NIBRS Data Analysis

  • There were more than 4.5 million arrests for alcohol-impaired driving reported to the FBI between 1995 and 2014, accounting for more than 10 percent of all reported arrests.
  • Alcohol-impaired driving arrest rates varied 32-fold across states.
  • The population-based rate of alcohol-impaired driving arrests reported to the FBI in 2014 decreased 13 percent from a high in 2003.
  • There has been a decrease in alcohol-impaired driving arrests as a proportion of all arrest activity in the United States.
  • There has been a notable and sustained increase in the percentage of female alcohol-impaired driving arrestees over the past 20 years.

There were 4,681,475 recorded DWI arrests reported to the NIBRS over the years 1995 to 2014, which accounted for 10.2 percent of all 45,978,155 reported arrests during that period. The overall population-based rate of DWI arrests for the entire 20-year period was 526.5 arrests per 100,000 persons covered by the reporting agencies. The average age of an arrestee was 34.4 years old (95% CI 34.3, 34.4), and listed ages ranged from 7 to 99 years old. Approximately 78.0 percent (95% CI 78.0, 78.1) of arrestees were male.

The median annual arrest rate for alcohol-impaired driving was 503.4 per 100,000 covered population. While a linear increasing trend line could be fit to the annual arrest rates, the data were more consistent with a period of increased enforcement from a low of 458.1 arrests per 100,000 covered population in 2003 to a high of 563.6 in 2009, followed by a 12.7 percent decline (95% CI 10.2, 15.8) to 491.9 in 2014.

When examined as the percentage of all arrests, there was an apparent decline in alcohol-impaired driving arrests as a proportion of all enforcement activity from 10 percent in 1995 to 9 percent in 2005, with evidence of some recovery in activity to approximately 10 percent of all arrests in 2014. The overall trend, though, indicated a significant decrease in alcohol-impaired driving arrests as a proportion of all arrest activity in the United States.

There was a slight increase in the mean age of arrestees for alcohol-impaired driving from approximately 34 years old in 1995 to 35.8 years old in 2014, with much of the increase occurring after 2010. Women were a rapidly increasing proportion of persons arrested for alcohol-impaired driving during the study period, increasing 40.0 percent (95% CI 22.0, 60.1) from 15.32 percent of all arrestees in 1995 to 25.56 percent in 2014 (see Figure A-4). There was a significant linear increase in the percentage of female arrestees of approximately 0.6 percent each year.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
images
FIGURE A-4 Proportion of women arrestees with overlying LOESS line, incidents of driving under the influence of alcohol.
SOURCE: NIBRS data, 1995–2014.

Rates of reported arrests for alcohol-impaired driving varied considerably by state from a low of 30.9 per 100,000 covered population in Delaware to a high of 960.8 in Kansas, with most states clustered around a median of approximately 490 arrests per 100,000 covered population and some evidence of regional variation. The percentage of all arrests represented by alcohol-impaired driving also varied across states, ranging from a low of 0.43 per 100,000 covered population in the District of Columbia to 21.87 in Vermont but with less overall heterogeneity. Persons arrested for alcohol-impaired driving were on average 5 years younger than the state median for every state except Delaware and Utah, where they were 2 to 3 years older. States also varied considerably in the percentage of women among alcohol-impaired driving arrestees, from a low of 6.98 percent of all arrestees in Georgia to 29.37 percent in Maine.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

NHTS Analysis

The NHTS estimates that the overall number of drivers increased 20 percent from 176,331,000 in 1995 to 212,309,000 in 2009, although growth was slower between 2001 and 2009 than between 1995 and 2001. The overall rate of growth varied by age group, and in fact there was an estimated decline in drivers in the 21- to 25-year-old and 26- to 35-year-old age groups (both men and women) between 1995 and 2009. By race and ethnicity, Hispanic drivers experienced the largest rate of growth, approximately 75 percent over the time period, followed by Asian drivers (66 percent), non-Hispanic blacks (33 percent), other non-Hispanic drivers (21 percent), and non-Hispanic whites (11 percent).

Estimates of total vehicle miles driven increased between 1995 and 2001 and decreased from 2001 to 2009, for a net change of +8.8 percent overall from 2.063 trillion in 1995 to 2.245 trillion in 2009. Patterns by age corresponded with driver trends; that is, there were declines in vehicle miles driven for adults 21 to 35 years of age. Vehicle miles driven by race and ethnicity mirrored the aforementioned rates of growth among drivers by race and ethnicity.

Estimated total miles per person of travel across all modalities (private vehicle, walking, biking, transit, air, and other) likewise increased between 1995 and 2001 and decreased between 2001 and 2009, with a net increase of 9 percent from 3.411 trillion to 3.732 trillion. By modality, private vehicle miles accounted for the majority of travel, increasing 6 percent from 3.110 trillion in 1995 to 3.298 trillion in 2009. A decline in overall vehicular travel was observed among 25- to 34-year-olds, similar to the observations above for drivers and vehicle miles driven. Men 16 to 24 years of age also experienced less overall vehicular travel from 1995 to 2009. While a smaller fraction of total travel, walking and biking estimates meanwhile increased both between 1995 and 2001 and between 2001 and 2009. Walking miles increased remarkably by 160 percent from 10.821 billion in 1995 to 27.943 billion in 2009. Estimated total biking miles likewise grew sharply by 95 percent from 4.585 billion in 1995 to 8.956 billion in 2009. Increases in miles walked were observed across all categories of age and gender. Biking trends by age and gender had some fluctuations likely as a result of smaller numbers.

Based on our review of the FHWA HVMT, across the United States, vehicle miles traveled increased between 1995 and 2008, declined somewhat between 2007 and 2009, and then increased again between 2010 and 2015.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

FARS Data Analysis

  • From 1995 to 2015 there was a 30 percent decrease in the absolute number of alcohol-related motor vehicle crash deaths in the United States, while there was a population increase of 40 percent and a 35 percent increase in VMT.
  • The greatest declines in alcohol-related motor vehicle crash mortality occurred among the youngest drivers and passengers.
  • Men accounted for 77 percent of all alcohol-related motor vehicle crash deaths in the United States.
  • Approximately 60 percent of alcohol-related motor vehicle crash deaths in the United States occur in rural areas.
  • Adjusted for VMT, the risk of an alcohol-related motor vehicle crash fatality in rural areas is nearly three times that of urban areas.
  • Non-Hispanic Asians had population-adjusted alcohol-related motor vehicle crash fatality rates more than three times lower than any other racial and ethnic group.
  • Motorcyclists, particularly those 55 to 64 years old, experienced a marked increase in alcohol-related crash fatalities.

This paper analyzes the FARS data from the years 1995 to 2015, acquired online from NHTSA. The FARS is a nationwide U.S. census providing annual data on all fatalities suffered from MVCs. Specifically, this study examines MVC fatalities with documented alcohol use. Fatal car crashes were considered alcohol related if any motor vehicle driver had police-reported alcohol intoxication or a positive laboratory BAC (BAC = 0.01 g/dL). Alcohol use by passengers, pedestrians, and cyclists did not classify a crash as an alcohol-related MVC.

Analyses focused on persons fatally injured from these crashes. Data were analyzed over this 21-year interval in terms of age, sex, race, and ethnicity. In addition, analyses compared urban and rural crash rates as well as crash rates between different states. Subanalyses were also performed on fatally injured cyclists, pedestrians, and motor vehicle occupants. Crash and fatality rates were adjusted in terms of census population counts obtained from the HCUP website as well as estimates of person-miles traveled obtained from the NHTS website. Results are presented as tables and time series plots.3 Analyses were completed using R.

___________________

3 Available by request from the National Academies of Sciences, Engineering, and Medicine’s Public Access Records Office (PARO@nas.edu).

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

Throughout the 1995–2015 study period, deaths from alcohol-related MVCs decreased by 29 percent. This decline was even greater after adjusting for population increases (41 percent) and person-miles traveled (37 percent). The greatest reductions were observed in younger persons, with 16- to 20-year-olds and children under 16 experiencing 56 and 66 percent decreases in alcohol-related MVC fatalities, respectively. The age group most frequently fatally injured in alcohol-related MVCs, 21- to 24-year-olds, experienced a 34 percent decrease in fatalities from 2,118 deaths in 1995 to 1,393 in 2015. Interestingly, individuals from 45 to 64 years of age experienced a 30 percent increase in alcohol-related MVC fatalities over the study period, but after adjusting for population increases this group also experienced a modest decline (15 percent).

Men accounted for 77 percent of all alcohol-related MVC fatalities. Despite this, both men and women experienced a decline in alcohol-related MVC fatalities of similar proportion, with male fatalities decreasing 29 percent from 10,667 deaths in 1995 to 7,545 in 2015, and female fatalities also decreasing 29 percent from 3,270 in 1995 to 2,312 in 2015. Non-Hispanic whites accounted for 68 percent of alcohol-related MVC fatalities from 2000 to 2015. However, after accounting for population size, non-Hispanic whites had approximately equal alcohol-related MVC fatalities per 100,000 population (3.52) as Hispanics (3.41) and non-Hispanic blacks (3.44) from 2000 to 2015. Non-Hispanic Asians, on the other hand, had population-adjusted fatality rates more than three times lower than all other racial and ethnic groups (1.08). During this period, non-Hispanic whites experienced a 24 percent decline in population-adjusted, alcohol-related MVC fatalities, compared to declines of 10, 56, and 41 percent in non-Hispanic blacks, non-Hispanic Asians, and Hispanics, respectively.

From 1995 to 2014, 60 percent of alcohol-related MVC fatalities occurred in rural areas. After adjusting for VMT, rural areas had a fatality rate of 7.24 deaths per billion VMT, 2.67 times greater than that of urban areas (2.71 deaths per billion VMT). This was relatively constant over the study period (see Figure A-5). During this period rural areas also experienced a greater decline in alcohol-related MVC fatalities (41 percent) compared to urban areas. However, after adjusting for VMT, the declines in MVC fatalities of urban areas (38 percent) and rural areas (40 percent) were similar. Of note, declines in rural fatalities were disproportionately reduced in individuals under 45 years of age (52 percent decrease), with the largest decreases in those under 21 years of age (64 percent). Rural individuals 45 and older had a slight increase in fatalities (4 percent). Changes in urban fatalities had less variance by age between individuals under 45 years of age (22 percent decline) and 45 years of age or older (22 percent increase).

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
images
FIGURE A-5 Deaths from alcohol-related crashes by year per billion vehicle miles traveled—urban versus rural.
SOURCE: FARS data, 1995−2014.

Of all those fatally injured in alcohol-related motor vehicle crashes, 95.4 percent were occupants of motor vehicles, 4.0 percent were pedestrians, and 0.6 percent were cyclists. After adjusting for person-miles traveled, pedestrians had the highest fatality rate (22.2 fatalities per billion person-miles traveled), followed by cyclists (11.2) and then motorists (5.2). Over the study period, pedestrians had the greatest reduction in fatalities (36 percent) compared to motor vehicle occupants (29 percent) and cyclists (22 percent). Motorcyclists, particularly those 55 to 64 years old, experienced a marked increase in both number and rate of alcohol-impaired fatalities, with a specific increase following 2010.

From 1995 to 2015 the yearly rate of alcohol-related MVC fatalities was 4.44 persons per 100,000 population. Fatality rates differed substantially between states with the highest population-adjusted fatality rates in Montana, New Mexico, North Dakota, South Carolina, and Wyoming (8.13–11.38 yearly deaths per 100,000 population) and the lowest fatality rates in the District of Columbia, Massachusetts, New Jersey, New York, and Utah (1.73–2.38). Four states, Maine, North Dakota, Rhode Island, and South Dakota, experienced increases in population-adjusted fatalities from 1995 to 2015 (from 1 to 28 percent), while the District of Columbia, Hawaii, Kansas, New Mexico, and Vermont had the greatest declines in population-adjusted fatality rates, with reductions ranging from 58 to 69 percent.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

Limitations

In this FARS database analysis, MVCs were considered alcohol related if a motor vehicle driver had police-reported alcohol intoxication or a positive laboratory BAC (BAC greater than 0.01 g/dL). Specifically, this information was obtained from the DR_DRINK variable from the FARS vehicle data files. Of note, the FARS crash-level DRUNK_DR variable was miscoded in the database for the years 1999–2007. While every crash within the database with driver alcohol involvement was included in this study, alcohol data are often missing from the database, meaning that this analysis may undercount the actual number of alcohol-related fatalities.

While the entire study period ranged from 1995 to 2015, analyses of race and ethnicity were from the years 2000 to 2015, since the FARS database did not begin recording race until 1999, and the U.S. Census significantly modified how it recorded race and ethnicity starting in 2000. As a result, the year 2000 was selected as the starting point for race and ethnicity subanalyses. Importantly, because the FARS obtains race and ethnicity information from death certificates, data are only available for individuals who were fatally injured. Even in this group, race or ethnicity information was missing in 14 percent of those persons reported to have died. This led to reduced fatality rates in the study’s analyses by race and ethnicity. Though the FARS had a small degree of missing data for other data elements, none of the absent data significantly affected analyses since the information was missing in well below 1 percent of persons fatally injured.

The FARS database is updated every year with several new and recoded variables, and this recoding did limit the urban and rural sub-analysis to a minor extent. While rural and urban designation was previously a dichotomous variable, in 2015, the FARS recoded its road classification data element to include a third interstate category. Since this led to reduced rural and urban crash rates, 2015 data were excluded and only the years 1995–2014 were examined. In addition, because the HCUP census uses a different criterion to distinguish rural and urban populations from the FARS, only adjusted fatality rates by VMT could be obtained for these urban and rural subanalyses.

REFERENCES

AHRQ (Agency for Healthcare Research and Quality). 2015. Introduction to the Healthcare Cost and Utilization Project’s (HCUP) Nationwide Emergency Department Sample (NEDS), 2013. https://www.hcup-us.ahrq.gov/db/nation/neds/NEDS2013Introduction.pdf (accessed October 24, 2017).

Brady, J. E., S. P. Baker, C. Dimaggio, M. L. McCarthy, G. W. Rebok, and G. Li. 2009. Effectiveness of mandatory alcohol testing programs in reducing alcohol involvement in fatal motor carrier crashes. American Journal of Epidemiology 170(6):775–782.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

Clark, D. E., K. L. Anderson, and D. R. Hahn. 2004. Evaluating an inclusive trauma system using linked population-based data. Journal of Trauma-Injury, Infection, and Critical Care 57(3):501−509.

FBI (Federal Bureau of Investigation). 2016. FBI releases 2015 crime statistics from the national incident-based reporting system, encourages transition. Washington, DC: FBI National Press Office. https://www.fbi.gov/news/pressrel/press-releases/fbi-releases-2015-crime-statistics-from-the-national-incident-based-reporting-system-encourages-transition (accessed March 7, 2018).

Loo, B. P., and K. Tsui. 2007. Factors affecting the likelihood of reporting road crashes resulting in medical treatment to the police. Injury Prevention 13(3):186−189.

Lopez, D. G., D. L. Rosman, G. A. Jelinek, G. J. Wilkes, and P. C. Sprivulis. 2000. Complementing police road-crash records with trauma registry data—An initial evaluation. Accident Analysis & Prevention 32(6):771−777.

Lujic, S., C. Finch, S. Boufous, A. Hayen, and W. Dunsmuir. 2008. How comparable are road traffic crash cases in hospital admissions data and police records? An examination of data linkage rates. Australian and New Zealand Journal of Public Health 32(1):28–33.

Miller, J. D., I. Cisin, and A. Harrel. 1986. A new technique for surveying deviant behavior: Item-count estimates of marijuana, cocaine, and heroin. Paper presented at the Annual Meeting of the American Association for Public Opinion Research, St. Petersburg, FL.

Schroeder, T., and K. Ault. 2001. The NEISS sample (design and implementation) 1997 to present. https://www.cpsc.gov/s3fs-public/pdfs/blk_media_2001d011-6b6.pdf (accessed October 24, 2017).

Sterne, J. A., I. R. White, J. B. Carlin, M. Spratt, P. Royston, M. G. Kenward, A. M. Wood, and J. R. Carpenter. 2009. Multiple imputation for missing data in epidemiological and clinical research: Potential and pitfalls. BMJ 338:B2393.

Subramanian, R. 2002. Transitioning to multiple imputation—A new method to estimate missing blood alcohol concentration (BAC) values in FARS. Washington, DC: National Center for Statistics and Analysis.

Subramanian, R., and D. Utter. 2003. Multiple imputation of missing blood alcohol concentration (BAC) values in FARS. Paper presented at the Federal Committee on Statistical Methodology Conference, Arlington, VA.

Warner, S. L. 1965. Randomized response: A survey technique for eliminating evasive answer bias. Journal of the American Statistical Association 60(309):63−69.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×

This page intentionally left blank.

Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 419
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 420
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 421
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 422
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 423
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 424
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 425
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 426
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 427
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 428
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 429
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 430
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 431
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 432
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 433
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 434
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 435
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 436
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 437
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 438
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 439
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 440
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 441
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 442
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 443
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 444
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 445
Suggested Citation:"Appendix A: Alcohol-Impaired Driving in the United States: Review of Data Sources and Analyses - Charles DiMaggio, Katherine Wheeler-Martin, and Jamie Oliver ." National Academies of Sciences, Engineering, and Medicine. 2018. Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem. Washington, DC: The National Academies Press. doi: 10.17226/24951.
×
Page 446
Next: Appendix B: Content Analysis of Alcohol-Impaired Driving Stories in the News - Deborah A. Fisher »
Getting to Zero Alcohol-Impaired Driving Fatalities: A Comprehensive Approach to a Persistent Problem Get This Book
×
Buy Paperback | $85.00 Buy Ebook | $69.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Alcohol-impaired driving is an important health and social issue as it remains a major risk to Americans' health today, surpassing deaths per year of certain cancers, HIV/AIDS, and drownings, among others, and contributing to long-term disabilities from head and spinal injuries. Progress has been made over the past decades towards reducing these trends, but that progress has been incremental and has stagnated more recently.

Getting to Zero Alcohol-Impaired Driving Fatalities examines which interventions (programs, systems, and policies) are most promising to prevent injuries and death from alcohol-impaired driving, the barriers to action and approaches to overcome them, and which interventions need to be changed or adopted. This report makes broad-reaching recommendations that will serve as a blueprint for the nation to accelerate the progress in reducing alcohol-impaired driving fatalities.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!