Content uploaded by Kathryn Wochinger
Author content
All content in this area was uploaded by Kathryn Wochinger on Apr 10, 2015
Content may be subject to copyright.
NHTSA’s Office of Behavioral Safety Research 1200 New Jersey Avenue SE., Washington, DC 20590
TRAFFIC SAFETY FACTS
Research Note
Office of Behavioral Safety ResearchDOT HS 812 118 February 2015
of drivers using alcohol or drugs and adjusted with an
appropriate weighting scheme.
Summary of Results
Prevalence of Alcohol Use by Drivers
The NRS surveys reveal a decreasing trend in alcohol use
from the rst survey in 1973 to the most recent one in 2013–
2014. Figure 1 shows the percentage of weekend nighttime
drivers with BrACs across three categories: BrAC of .005 to
.049 g/210 L; 2 BrACs of .050 to .079; and BrACs of .080 and
higher. The surveys found a decline in each BrAC category.
Further, there has been a large decrease in the percentage of
drivers who were alcohol positive, from 35.9 percent in 1973
to 8.3 percent in 2013–2014. For BrACs of .08 and higher,
there was a decrease from 7.5 percent in 1973 to 1.5 percent
in 2013–2014, revealing an impressive 80 percent reduction
in the percentage of alcohol-impaired drivers on the road
on weekend nights. Also of importance is the decrease
Over the last four decades, the National Highway Trafc
Safety Administration and/or the Insurance Institute for
Highway Safety (IIHS) conducted four national surveys
to estimate the prevalence of drinking and driving in the
United States (Wolfe, 1974; Lund & Wolfe, 1991; Voas et al,
1998; Compton & Berning, 2009; Lacey et al, 2009). The rst
National Roadside Survey (NRS) was conducted in 1973,
followed by national surveys of drivers in 1986, 1996, 2007,
and now 2013–2014. These surveys used a stratied random
sample of weekend nighttime drivers in the contiguous 48
States and collected data directly from drivers on the road.
The 2007 NRS added procedures to the NRS for the rst
time to estimate the use by drivers of other potentially
impairing drugs. Prior roadside surveys had only collected
breath samples to determine breath alcohol concentration
(BrAC). Due to developments in analytical toxicology,
NHTSA determined it would be feasible in the 2007 and
2013–2014 surveys to determine driver use of a variety of
potentially impairing drugs including illegal drugs as well
as legal medications.
In 2013–2014, the National Highway Trafc Safety
Administration conducted the most recent National
Roadside Survey of Alcohol and Drug Use by Drivers.1
This voluntary and anonymous study is the second to col-
lect data on drug use, presenting our rst opportunity to
examine drug use trends on a national scale. The 2013–
2014 NRS was designed to produce national estimates of
alcohol and drug use by weekday daytime and weekend
nighttime drivers. Thus, the use rates presented below are
national prevalence rates calculated from the percentage
Results of the 2013–2014 National Roadside
Survey of Alcohol and Drug Use by Drivers
By Amy Berning, Richard Compton, and Kathryn Wochinger
1 The Ofce of National Drug Control Policy provided funds to
NHTSA for this study. The National Institute on Drug Abuse, and
the Insurance Institute for Highway Safety provided additional
funding through investigator initiated grants and contracts to the
Pacic Institute for Research and Evaluation, NHTSA’s contractor
for the 2013–2014 NRS.
2 .08 g/210 L = grams per 210 liters of breath. The illegal limit in all
States is .08.
3 From 1973 to 2004, the States had BrAC limits that ranged from .08
to .15. After 2004, all States had BrAC limits of .08.
Figure 1.
Percentage of Weekend Nighttime Drivers by BrAC
Category in the Five National Roadside Surveys3
0
5
10
15
20
25
19861973
BrAC .005–.049
BrAC .050–.079
BrAC .08+
1996 2007 2013–2014
22.3
6.17.5
17.6
3
5.4
9.2
3.44.3
7.9
2.32.2
5.2
1.6 1.5
Percentage of Drivers
2
NHTSA’s Office of Behavioral Safety Research 1200 New Jersey Avenue SE., Washington, DC 20590
from 6.1 percent to 1.6 percent from 1973 to 2013–2014 for
BrACs of .050 to .079 category.
The 2013–2014 survey found large differences by the day of
week and the time of day in the likelihood of drivers being
alcohol positive or having an illegal BrAC (Table 1). During
weekday daytime hours (Friday), only 1.1 percent of drivers
were alcohol positive, while at weekend nighttime hours
(Friday and Saturday), 8.3 percent of drivers were alcohol
positive. During weekday daytime hours there were very
few drivers with illegal BrACs (BrAC ≥ .08), just 0.4 percent,
while at weekend nighttime hours 1.5 percent drivers had
illegal BrACs. Daytime compared to nighttime percentages
are statistically signicant. Compared with the NRS 2007,
the 2013–2014 NRS BrAC prevalence shown in Figure 1 was
signicantly lower only for the .005 to .049 BrAC category.
Table 1
Alcohol Prevalence by Data Collection Period and BrAC in
the 2013–2014 NRS
Data Collection Time Period
% Alcohol Positive
(%BrAC > .005) % BrAC > .08
Weekday Daytime 1.1% 0.4%
Weekend Nighttime 8.3% 1.5%
Prevalence of Drug Use by Drivers
The 2013–2014 study examined the use of drugs, focus-
ing on drugs with the potential to impair driving skills,
including over-the-counter, prescription, and illegal drugs.
Participants were asked to provide an oral uid and blood
sample in addition to a breath sample. The oral uid and
blood samples were tested for the presence of a large num-
ber of potentially impairing drugs including cannabinoids,
stimulants, sedatives, antidepressants, and narcotic analge-
sics. Not all drivers provided both an oral uid and blood
sample; some drivers provided just one sample but many
provided both.
The reader is cautioned that drug presence does not neces-
sarily imply impairment. For many drug substances, drug
presence can be detected after impairment that might affect
driving has passed. For example, traces of marijuana use
can be detected in blood samples several weeks after heavy
chronic users stop ingestion. In this study, for marijuana,
we tested only for THC (delta 9 tetrahydrocannabinol), the
psychoactive substance in marijuana, and 11-OH-THC, its
active metabolite. When marijuana is smoked or ingested,
THC is absorbed into the blood stream and is distributed
into areas of the body, including the brain. There are over
100 marijuana metabolites detectable in blood that research
has not associated with the psychoactive effects of mari-
juana use. Whereas the impairment effects for various con-
centration levels of alcohol in the blood or breath are well
understood, there is little evidence available to link concen-
trations of other drugs to driver performance.
Table 2
Overall Drug Prevalence by Data Collection Period and Type
of Test in the 2013–2014 NRS
Time of Day
% Drug-Positive
Oral Fluid Test
% Drug-Positive
Blood Test
% Drug-Positive Oral
Fluid and/or Blood Test
Weekday
Daytime 19.0% 21.6% 22.4%
Weekend
Nighttime 19.8% 21.2% 22.5%
In contrast to alcohol use, overall drug prevalence (shown
in Table 2), did not appear to differ between daytime and
nighttime, regardless of whether oral uid or blood tests
were conducted. The much higher nighttime use of alco-
hol appears to represent recreational use. For other drugs, a
different pattern emerges.
As shown in Table 3, the pattern of drug use among driv-
ers varies by day and category of drug. The prevalence of
illegal drug use increases from daytime to nighttime, but
there is a countervailing pattern of a reduction in preva-
lence of driver use of medicinal drugs in nighttime drivers
compared to daytime drivers.4
Table 3
Drug Prevalence by Data Collection Period, Drug Category,
and Type of Drug Test in the 2013–2014 NRS
Drug Category
Oral Fluid
Test Blood Test
Oral Fluid and/or
Blood Test
N % N % N %
Weekday Daytime
Any Illegal Drug5189 10.6% 137 11.3% 221 12.1%
Only Medications6
( prescription and over-the-counter)
197 8.4% 128 10.3% 234 10.3%
Weekend Nighttime
Any Illegal Drug 783 13.9% 423 14.3% 852 15.2%
Only Medications
( prescription and over-the-counter)
317 5.9% 216 6.9% 391 7.3%
4 Despite recent changes in the legal status of marijuana in some
States, for simplicity and to allow inter-survey comparisons, this
drug remained included within the “illegal” category in the 2013–
2014 N R S.
5 “Any Illegal Drug” includes drivers who tested positive for one
or more illegal drugs, whether or not they also tested positive for
medications.
6 “Only Medications” includes drivers who tested positive for one
or more medications, but did not test positive for illegal drugs.
3
NHTSA’s Office of Behavioral Safety Research 1200 New Jersey Avenue SE., Washington, DC 20590
Comparison of Drug Prevalence Between the 2007
NRS and 2013–2014 NRS
The specic drugs and detection thresholds changed
somewhat between the 2007 NRS and 2013–2014 NRS. A
few drugs that were either not detected or rarely detected
in the 2007 NRS were not included on the 2013–2014 NRS,
and a few new drugs were added in the 2013–2014 NRS.
For example, we tested for selected synthetic cannabinoids
in the new survey. Also, improvements in analytical tech-
nology allowed the detection thresholds to be lowered
signicantly for a number of drugs in the 2013–2014 NRS.
To compare the prevalence rates between the 2007 NRS
and 2013–2014 NRS, we included only the drugs that were
tested for in both surveys and based 2013–2014 NRS preva-
lence rates using the cutoff levels from the 2007 NRS.
To account for these renements in lab testing since 2007,
Table 4 presents the 2013–2014 NRS data after applying the
same set of drugs and drug cutoff levels as used for 2007
survey. The results from either oral uid tests or blood
tests, or both oral uid and/or blood tests, show a small
increase in drug-positive drivers using medications (from
3.9% in 2007 to 4.9% in 2013–2014), and a larger increase in
the prevalence of illegal drugs (from 12.4% in 2007 to 15.1%
in 2013–2014). In 2007, 16.3 percent of weekend nighttime
drivers were drug-positive based on the combined results
of either or both oral uid and blood tests (Compton &
Berning, 2009). In 2013–2014, 20.0 percent of weekend
nighttime drivers tested positive for drugs using the 2007
cut-off criteria.
The drug with the largest increase in weekend night-
time prevalence was THC (Table 5). In the 2007 NRS, 8.6
percent of weekend nighttime drivers tested positive for
THC (based on the combined oral uid and/or blood tests),
whereas in the 2013–2014 NRS, 12.6 percent of weekend
nighttime drivers tested positive for THC, a 48 percent
increase.
Changes in State policy on marijuana use, including
medical and recreational use, may have contributed to an
increase in marijuana use by drivers. However, the survey
does not permit a state-by-state comparison. The change in
use may reect the emergence of a new trend in the coun-
try that warrants monitoring.
Survey Methodology
This most recent NRS started in the summer of 2013, contin-
ued through the spring of 2014, and involved 60 sites across
the contiguous United States. The study used a multistage
sampling procedure based on the National Automotive
Sampling System—General Estimates System (NHTSA,
2008). The sites were large cities, counties, or groups of
counties representing four regions within the United States
and three levels of population density. As a whole, the 60
sites provide a representative sample of drivers across the
country. The 2013–2014 NRS attempted to re-visit the same
sites as 2007; however, because not all sites were avail-
able, statistical sampling techniques were used to identify
replacement sites with similar characteristics.
Roadside survey data were collected in 5 different loca-
tions at each site (for a total of 300 locations). Although
there were practical considerations for selecting a site, such
as trafc ow and safety, the 5 locations were randomly
chosen within the boundaries of the cooperating local law
enforcement agency’s area of jurisdiction. This approach
provided, as much as possible, a representative sample
of drivers for that site. Locations were not selected on the
basis of assumptions of where there would be a high per-
Table 4
Weekend Nighttime Drug Prevalence by Drug Category and Test Type Comparing 2007 Data to 2013–2014 Comparable Data
Drug Category
2007 Data 2013–2014 Comparable Data
Oral Fluid Test Blood Test
Oral Fluid and/or
Blood Test Oral Fluid Test Blood Test
Oral Fluid and/or
Blood Test
N%N%N%N%N%N%
Any Illegal Drug 635 11.4% 297 9.8% 699 12.4% 779 13.8% 422 14.3% 849 15.1%
Only Medications
(prescription and over-the-counter)
201 3.0% 169 4.0% 277 3.9% 211 3.9% 155 4.9% 266 4.9%
Table 5
Weekend Nighttime Prevalence of THC in 2007 Compared to 2013–2014 Comparable Data
2007 Data 2013–2014 Comparable Data
Oral Fluid Test Blood Test Oral Fluid and/or Blood Test Oral Fluid Test Blood Test Oral Fluid and/or Blood Test
N % N % N % N % N % N %
438 7.7% 234 7.6% 499 8.6% 597 11.3% 332 11.7% 663 12.6%
4
NHTSA’s Office of Behavioral Safety Research 1200 New Jersey Avenue SE., Washington, DC 20590
centage of alcohol- or drug-positive drivers. The objective
of this study was to estimate the alcohol and other drug
prevalence of all drivers “on the road in the United States”
during the given time periods.
In the 2007 NRS, in an effort to learn more about the
presence of alcohol and drug use during daytime driv-
ing, we added data collection on Fridays during the day,
from either 9:30 to 11:30 a.m., or 1:30 to 3:30 p.m. (the times
were randomized across sites). This additional data collec-
tion allowed us to examine drug use and alcohol trends
among daytime drivers. This method was repeated for
the 2013–2014 NRS, and data were collected on weekend
nights, from 10:00 p.m. to midnight, and from 1 to 3 a.m. on
Fridays (i.e., early Saturday mornings) and from 1 to 3 a.m.
on Saturdays (i.e., early Sunday mornings). By using these
same time frames for each NRS, we can compare data on
BrACs and drug use over time.
Participation in the survey was entirely voluntary and
anonymous. When a data collector was available, drivers
were allowed to enter the data collection location. Where
researchers described the project and offered incentives
for participation. Drivers were informed that participation
was voluntary and that they were free to leave at any time.
Data collectors asked participants for a breath test, an oral
uid sample, and a blood sample. The vast majority of eli-
gible drivers entering the research area participated in the
study (see Table 6). The number of drivers who entered the
site and were eligible to participate (e.g., non-commercial
drivers 16 and older, English- or Spanish-speaking) was
11,100. Out of the 11,100 eligible drivers, 85.2 percent (9,455
drivers) provided breath samples; 71 percent (7,881 drivers)
provided oral uid samples; and 42.2 percent (4,686 drivers)
provided blood samples.
National prevalence rates were derived from a complex
weighting scheme based on the volume of serious crashes
at each site and the probability of a survey driver being ran-
domly selected from the total driving trips at that site.
Table 6
Number of Participants and Participation Rates in the
2013–2014 NRS
Drivers
Participants
N %
Eligible and Entered Site 11,100 100.0%
Breath Test 9,455 85.2%
Oral Fluid Test 7,881 71.0%
Blood Test 4,686 42.2%
Oral Fluid and/or Blood 7,898 71.2%
Challenges in Determining How Drugs
AffectDriving
Most psychoactive drugs are chemically complex mol-
ecules, whose absorption, action, and elimination from
the body are difcult to predict, and considerable differ-
ences exist between individuals with regard to the rates
with which these processes occur. Alcohol, in comparison,
is more predictable. A strong relationship between alco-
hol concentration and impairment has been established,
as has the correlation between alcohol concentration and
crashrisk.
Factors that make similar prediction difcult for most other
psychoactive drugs include:
■■The large number of different drugs that would need to
be tested (extensive testing of alcohol has been under-
taken over many decades, whereas relatively little simi-
lar testing has occurred for most other drugs).
■■Poor correlation between the effects on psychomotor,
behavioral, and/or executive functions and blood or
plasma drug concentrations (peak psychomotor, behav-
ioral, and executive function effects do not necessarily
correspond to peak blood levels; detectable blood levels
may persist beyond the impairing effects or the impair-
ing effects may be measurable when the drug cannot be
detected in the blood).
■■Sensitivity and tolerance (accentuation and diminution
of the impairing effects with repeated exposure).
■■Individual differences in absorption, distribution, action,
and metabolism (some individuals will show evidence
of impairment at drug concentrations that are not asso-
ciated with impairment in others; wide ranges of drug
concentrations in different individuals have been associ-
ated with equivalent levels of impairment).
■■Accumulation (blood levels of some drugs or their
metabolites may accumulate with repeated administra-
tions if the time-course of elimination is insufcient to
reduce or remove the drug or metabolite before the next
dose is administered).
■■Acute versus chronic administration (it is not unusual to
observe greater impairment during initial administra-
tions of drugs than is observed when the drug is admin-
istered over a long period of time).
At the current time, specic drug concentration levels can-
not be reliably equated with a specic degree of driver
impairment.
5
NHTSA’s National Center for Statistics and Analysis 1200 New Jersey Avenue SE., Washington, DC 20590
This and other behavioral-related research notes
may be accessed by Internet users at: www.nhtsa.
gov/Driving+Safety/Research+&+Evaluation/
Impaired+driving+(drug-related)+reports.
For More Information
For questions on this document, please contact Amy
Berning at amy.berning@dot.gov.
Detailed information about the study and results will
be available in upcoming publications. Three technical
reports are under development: one provides a complete
description of the methodology used (sampling, analy-
sis, weighting, and imputation procedures) and subject
participation rates (report entitled, 2013–2014 National
Roadside Survey of Alcohol and Drug Use: Methodology).
Detailed information on the use of alcohol by drivers and
the relationship of alcohol to various factors like time of
day and vehicle type will be available in a report entitled,
2013–2014 National Roadside Survey of Alcohol and Drug
Use: Alcohol Prevalence Rates. Detailed information on
the use of drugs by drivers and the combined use of drugs
and alcohol, as well as more detailed trend analysis exam-
ining changes in drug prevalence from 2007 to 2013–2014
will be provided in a third report titled 2013–2014 National
Roadside Survey of Alcohol and Drug Use: Drug Prevalence
Rates. These upcoming reports will be posted on NHTSA’s
Web site.
References
Compton, R., & Berning, A. (2009). Results of the 2007
National Roadside Survey of Alcohol and Drug Use by Drivers
(Report No. DOT HS 811 175). Washington, DC: National
Highway Trafc Safety Administration. Available at
www.nhtsa.gov/DOT/NHTSA/Traffic%20Injury%20
Control/Articles/Associated%20Files/811175.pdf.
Lacey, J. H., Kelley-Baker, T., Furr-Holden, D., Voas, R.
B., Moore, C., Brainard, K., … & Berning, A. (2009,
December). 2007 National Roadside Survey of Alcohol and
Drug Use by Drivers: Methodology. (Report No. DOT HS
811 237). Washington, DC: National Highway Trafc
Safety Administration. Available at www.nhtsa.gov/
staticles/nti/pdf/811237.pdf.
Lacey, J. H., Kelley-Baker, T., Furr-Holden, D., Voas, R.
B., Romano, E., Ramirez, A., … & Berning, A. (2009,
December). 2007 National Roadside Survey of Alcohol and
Drug Use by Drivers: Drug Results (Report No. DOT HS
811 249). Washington, DC: National Highway Trafc
Safety Administration. Available at www.nhtsa.gov/
DOT/NHTSA/Traffic%20Injury%20Control/Articles/
Associated%20Files/811249.pdf.
Lacey, J. H., Kelley-Baker, T., Furr-Holden, D., Voas, R.
B., Romano, E., Torres, P., … & Berning, A. (2009,
December). 2007 National Roadside Survey of Alcohol
and Drug Use by Drivers: Alcohol Results (Report No.
DOT HS 811 248). Washington, D.C: National Highway
Trafc Safety Administration. Available at ntl.bts.gov/
lib/31000/31600/316 43/811248.pdf.
Lund, A. K., & Wolfe, A. C. (1991). Changes in the incidence
of alcohol-impaired driving in the United States, 1973-
1986. Journal of Studies on Alcohol, 52(4), 293–301.
NHTSA. (2008). National Automotive Sampling System
Brochure (Report No. DOT HS 807 304). Washington,
D.C: National Highway Trafc Safety Administration.
Available at www.nhtsa.gov/DOT/NHTSA/NCSA/
Content/PDF/NASSbrochure.pdf.
Voas, R. B., Wells, J., Lestina, D., Williams, A., & Greene, M.
(1998). Drinking and driving in the United States: The
1996 National Roadside Survey. Accident Analysis and
Prevention, 30(2), 267–275.
Wolfe, A. C. (1974). 1973 US National Roadside Breath Testing
Survey: Procedures and Results. Ann Arbor, MI: University
of Michigan Safety Research Institute.
Suggested APA format citation for this document:
Berning, A., Compton, R., & Wochinger, K. (2015, February). Results
of the 2013–2014 National Roadside Survey of alcohol and drug
use by drivers. (Trafc Safety Facts Research Note. Report No.
DOT HS 812 118). Washington, DC: National Highway Trafc
Safety Administration.
11326-020515-v3
