Introduction
Controlled Substance Testing
regulations were published by the Federal Highway Administration in 1988.1 Effective December 21, 1989 these rules required commercial motor
carriers engaged in interstate commerce employing 50 or more truck drivers to
begin random urine drug screens. On December 21, 1990 all other interstate
carriers became subject to the same requirements.
The random drug testing regulations are governed by 49 CFR part 40. These regulations apply to employers of interstate commercial truck drivers and owner operators. Drivers to be tested are selected randomly by the employer with the odds of being selected equal for every covered driver on every collection day. The frequency and interval of collections are at the discretion of the employer as long as 50% of the covered drivers are tested annually. Collection sites are designated by the employer and must posses the necessary personnel, facilities, and material to collect the specimens in accordance with protocols established in the rule. Specimens must be analyzed by a federally certified laboratory.
The authors undertook a literature review in search
of papers that have analyzed the effect this regulation has had on the rate of
commercial motor vehicle crashes. We performed a search of the peer reviewed as
well as government literature but could find no papers that have addressed this
question. It is important to understand what impact this regulatory burden has
had on achieving improved commercial motor vehicle safety.
Once it was determined that this question had not
been addressed in the published literature we began a search for the data
needed to undertake this analysis. We discovered that there was good numerator
data on the casualty side but very poor denominator data on the numbers of
active truck drivers. Further research revealed data on fatal accidents per 100
million vehicle miles traveled.2 This is the data the government uses as a proxy for accident rates.
The data showed that the fatal accident rate has
been on a downward trend during the years prior to and since the inception of
drug screening
Methods
This paper analyzes data on fatal truck crashes.
This analysis was initially undertaken by searching for incidence rates on
crashes among commercial motor vehicle drivers. After searching the federal
government databases and publications it became apparent that this data might
not exist in a form that is accurate and reliable. A representative of the
Analysis Division at the Federal Highway Administration indicated that accident
rates for truck drivers are not available. There is good numerator data on the
number of fatal large truck crashes each year, however, denominator data on the
number of active truck drivers is not known with a sufficient degree of
accuracy. The government uses rates of truck crashes per 100 million vehicle
miles traveled (VMT) as a proxy for crash rates among truck drivers. This data
is contained in the Fatality Analysis Reporting System database (FARS) and is
summarized in Traffic Safety Facts 1997.3,2
The General Estimates System (GES) is a database
that contains information on injury and property-damage-only crashes.4
This data goes back to 1988 which does not provide sufficient
pre-drug-screening data to allow meaningful analysis.
The change in the Fatality
Rate per 100 million VMT was modeled by linear regression over the years 1984
to 1989. (Fig. 1) In addition, 95% prediction limits were computed (95%
confidence intervals for predicting an individual value). These were
extrapolated over the following 8 years. Values that fell outside these limits
were considered significantly different from the trend exhibited over the years
1984-1989.
Results
The data on crash rates per
100 million vehicle miles traveled is taken from Traffic Safety Facts 1997.2
It is abstracted from the FARS database which is maintained by the National
Highway Traffic Safety Administration.3 (Table 1) A large truck is
defined by FARS as weighing over 10,000 pounds gross vehicle weight, including
single unit trucks and truck tractors.5
Discussion
We conclude from this analysis that the introduction
of random urine drug screening among commercial truck drivers was associated
with a statistically significant decrease in fatal truck crashes for the 2
years following widespread application of the regulations.
By December 21, 1990 all interstate commercial motor
vehicle carriers were subject to the random drug testing requirements. The
first full calendar year that all interstate large truck drivers were subject
to random testing under the regulations was 1991.The data points for 1991 and
1992 fall outside the confidence limits of the regression line. By 1993 the
fatal crash rate was back within the confidence limits of the regression line
and the data points for 1996 and 1997 fall on the regression line. (Fig. 1) The
authors could not find suitable data to use as a control.
According to an analyst with the Federal Highway
Administration, the FARS database while considered reliable, is subject to
error. Trucks can be miscoded thus introducing classification bias into the
data. The database includes all trucks, not just commercial trucks. As a
result, government and private non-commercial trucks are included. These
vehicles are not subject to random urine drug screening regulations and may
skew the data. Also, the VMT data has been subject to reporting error and there
have been variations from state-to-state and year-to-year. However, over the
years as data collection and review has become more consistent the Federal
government and industry have come to use this data as an important indicator of
commercial truck safety. We believe that the trends established by this data
are useful in spite of the potential sources of bias.
Fatal crashes are only part of the commercial truck
safety picture. The lack of pre random drug testing data on injury and
property-damage-only crashes together with a lack of suitable control data
limited our ability to make supportable conclusions.
A
representative of the Federal Motor Carrier Safety Administration within the
Department of Transportation was consulted to determine whether other
commercial motor vehicle safety regulations might have been effective over time
in contributing to the observed accident rate decrease. The Commercial Motor
Vehicle Safety Act of 1986 was the enabling legislation that required covered
drivers to obtain a Commercial Drivers License (CDL). These regulations (49 CFR
part 383) required drivers to obtain a CDL by April 1, 1992. It is unlikely
that this requirement had any impact on the accident rate for 1991. It is
possible that the 1992 rate was affected to a limited degree. Nineteen
ninety-three would have been the first full year affected by the CDL
requirement and by then the accident rate decrease was no longer statistically
significant.
Our analysis does not establish causation. We cannot
be certain that the initiation of random drug screening of truck drivers was
responsible for the temporary reduction of fatal truck crashes. It is likely
that drug screening was responsible for a portion of the observed decrease.
Random
drug screening may have produced benefits early on that do not appear to be
sustained in later years. Can the costs and regulatory burden of randomly drug
screening commercial truck drivers continue to be justified given this datas
apparent lack of continued benefit to public safety? This paper raises public
policy questions that need to be explored further. Hopefully this analysis will
stimulate debate and continued study of this issue.
Acknowledgments
The authors thank Mr. Mark Riggs for his advice and
help with statistical treatment of the data, Ralph Craft for technical
assistance with the truck crash data, and Larry Minor for assistance with
applicable truck safety regulations.
Research for this study was self- funded. No
financial assistance was received from any outside source.
1. Controlled
Substances Testing. (1988). Code of Federal Regulations [Print], Tit. 49, Pts.
391 and 394. Available: Federal Register [1988, November 21].
2. Traffic Safety Facts 1997: A Compilation of
Motor Vehicle Crash Data from the Fatality Analysis Reporting System and the
General Estimates System. Washington DC: U.S. Department of Transportation,
National Highway Traffic Safety Administration; 1998:17.
3. Fatality
Analysis Reporting System. National Center for Statistics and Analysis.
Available: http://www.nhtsa.dot.gov/people/ncsa/fars.html. [1999, August 9].
4.
General
Estimates System. National Center for Statistics and Analysis. Available:
http://www.nhtsa.dot.gov/people/ncsa/nass_ges.html. [1999, August 9].
5.
Traffic Safety
Facts 1997: 1998:197.
Fig.
1. Linear regression analysis of data from 1984 to 1989 is extrapolated over
the subsequent 8 years. 95% confidence intervals are shown.
Table 1
Large truck fatal accident
rate* |
|
|
||
Year
|
Accident Rate
|
|
||
|
1984 |
4.2 |
|
||
|
1985 |
4.2 |
|
||
|
1986 |
4.0 |
|
||
|
1987 |
3.8 |
|
||
|
1988 |
3.8 |
|
||
|
1989 |
3.5 |
|
||
|
1990 |
3.3 |
|
||
|
1991 |
2.9 |
|
||
|
1992 |
2.6 |
|
||
|
1993 |
2.7 |
|
||
|
1994 |
2.7 |
|
||
|
1995 |
2.5 |
|
||
|
1996 |
2.6 |
|
||
|
1997 |
2.5 |
|
||
|
|
|
|||
*per 100 million vehicle
miles traveled (VMT)