[Federal Register Volume 73, Number 224 (Wednesday, November 19, 2008)]
[Rules and Regulations]
[Pages 69567-69586]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: E8-27437]
=======================================================================
-----------------------------------------------------------------------
DEPARTMENT OF TRANSPORTATION
Federal Motor Carrier Safety Administration
49 CFR Parts 385 and 395
[Docket No. FMCSA-2004-19608]
RIN-2126-AB14
Hours of Service of Drivers
AGENCY: Federal Motor Carrier Safety Administration (FMCSA), DOT.
ACTION: Final rule.
-----------------------------------------------------------------------
SUMMARY: FMCSA adopts as final the provisions of the Agency's December
17, 2007, interim final rule concerning hours of service (HOS) for
commercial motor vehicle (CMV) drivers. This final rule allows CMV
drivers to continue to drive up to 11 hours within a 14-hour, non-
extendable window from the start of the workday, following at least 10
consecutive hours off duty (11-hour rule). The rule also allows motor
carriers and drivers to continue to restart calculations of the weekly
on-duty limits after the driver has at least 34 consecutive hours off
duty (34-hour restart).
DATES: Effective Date: This rule is effective January 19, 2009.
Docket: For access to the docket to read background documents or
comments received, go to http://www.regulations.gov at any time or to
the ground floor, room W12-140, U.S. Department of Transportation (DOT)
Building, 1200 New Jersey Avenue, SE., Washington, DC, between 9 a.m.
and 5 p.m. e.t., Monday through Friday, except Federal holidays.
Privacy Act: Anyone is able to search the electronic form of all
comments received into any of our dockets by the name of the individual
submitting the comment (or signing the comment, if submitted on behalf
of an association, business, labor union, etc.). You may review DOT's
complete Privacy Act Statement in the Federal Register published on
April 11, 2000 (65 FR 19476) or you may visit http://docketsinfo.dot.gov.
FOR FURTHER INFORMATION CONTACT: Mr. Thomas Yager, Chief, FMCSA Driver
and Carrier Operations. Telephone (202) 366-4325 or E-mail
[email protected].
SUPPLEMENTARY INFORMATION:
Table of Contents
A. Legal Basis for the Rulemaking
B. Background
C. Discussion of Rule
D. Discussion of Comments
1. Statutory Duty
2. Comments on Safety
3. Comments on Driver Health
4. Approach to Research
5. Use of Data, Analysis, and Modeling
E. Regulatory Analyses and Notices
F. List of References
G. Removal of Rescission Provision
A. Legal Basis for the Rulemaking
This rule is based on the authority of the Motor Carrier Act of
1935 and the Motor Carrier Safety Act of 1984. The Motor Carrier Act of
1935 provides that ``The Secretary of Transportation may prescribe
requirements for (1) qualifications and maximum hours of service of
employees of, and safety of operation and equipment of, a motor
carrier; and, (2) qualifications and maximum hours of service of
employees of, and standards of equipment of, a motor private carrier,
when needed to promote safety of operation'' (Section 31502(b) of Title
49 of the United States Code (49 U.S.C.)).
The HOS regulations adopted in this final rule concern the
``maximum hours of service of employees of * * * a motor carrier'' (49
U.S.C. 31502(b)(1)) and the ``maximum hours of service of employees of
* * * a motor private carrier'' (49 U.S.C. 31502(b)(2)). The adoption
and enforcement of such rules were specifically authorized by the Motor
Carrier Act of 1935. This rule rests on that authority.
The Motor Carrier Safety Act of 1984 provides concurrent authority
to regulate drivers, motor carriers, and vehicle equipment. It requires
the Secretary of Transportation to ``prescribe regulations on
commercial motor vehicle safety. The regulations shall prescribe
minimum safety standards for commercial motor vehicles.'' Although this
authority is very broad, the 1984 Act also includes specific
requirements: ``At a minimum, the regulations shall ensure that (1)
commercial motor vehicles are maintained, equipped, loaded, and
operated safely; (2) the responsibilities imposed on operators of
commercial motor vehicles do not impair their ability to operate the
vehicles safely; (3) the physical condition of operators of commercial
motor vehicles is adequate to enable them to operate the vehicles
safely; and (4) the operation of commercial motor vehicles does not
[[Page 69568]]
have a deleterious effect on the physical condition of the operators''
[49 U.S.C. 31136(a)].
This rule is also based on the authority of the 1984 Act and meets
the specific mandates of 49 U.S.C. 31136(a)(2), (3), and (4). Section
31136(a)(1) primarily governs the mechanical condition of CMVs, a
subject not included in this rulemaking. To the extent the phrase
``operated safely'' in paragraph (a)(1) encompasses safe driving, this
rule also addresses that mandate.
Before prescribing any regulations, FMCSA must also consider their
``costs and benefits'' (49 U.S.C. 31136(c)(2)(A) and 31502(d)). Those
factors are also discussed in this final rule.
B. Background
For background information on this rulemaking, please see the
account published in the interim final rule (IFR) of December 17, 2007
[72 FR 71247, 71250-71251].
C. Discussion of Rule
FMCSA is promulgating as a final rule the provisions of the IFR it
adopted on December 17, 2007. Because the United States Court of
Appeals for the District of Columbia Circuit (the Court or DC Circuit)
held in 2007 that the Agency had failed to provide an opportunity for
public comment on certain aspects of the 2005 Regulatory Impact
Analysis (RIA) [Owner-Operator Independent Drivers Association, Inc. v.
Federal Motor Carrier Safety Administration, 494 F.3d 188 (DC Cir.
2007)], the IFR provided a 60-day period for the public to comment on
the RIA. In response to the Court's finding that FMCSA did not provide
an adequate explanation for certain critical elements in one of its
analytical models used in the RIA, the preamble to the IFR also
included a detailed explanation of the Agency's time-on-task (TOT)
methodology [72 FR 71252 et seq.], thus satisfactorily addressing the
second flaw identified by the DC Circuit.
Most of the comments to the IFR docket reiterated arguments and
conclusions set forth during the 2003 and 2005 HOS rulemakings; the
more significant comments are discussed below.
Before addressing those comments, it is useful to summarize the
reasoning that led the Agency to adopt the 2005 HOS rule, which was
restored by the IFR and is finalized by today's action. Research on the
causes and effects of fatigue is sometimes inconsistent, frequently
based on work environments other than truck driving, and usually
conducted on a small scale. It is not unusual for an assertion or
conclusion related to fatigue to be questioned in some published study.
Researchers have also examined environmental factors related to many
potential driver health issues, but these studies are not sufficiently
precise to allow reasonable estimates of the benefits of remedial
measures.\1\
---------------------------------------------------------------------------
\1\ See the extended discussion in the preamble of the 2005
final rule, 70 FR 49982-49992.
---------------------------------------------------------------------------
Due to the lack of clear and consistent scientific evidence in this
area, the Agency went to great lengths to review the research
literature, utilizing resources of the Transportation Research Board of
the National Academies, and the National Institute for Occupational
Safety and Health (NIOSH), as well as experts from other DOT
organizations. FMCSA's own expertise and judgment were particularly
significant in reviewing, evaluating, and properly weighting research
findings. FMCSA's unique knowledge of the motor carrier industry and
its patterns of operation provide the Agency with a sound basis for
assessing the safety impact of this rulemaking action. In fact, FMCSA's
own field surveys, conducted in 2005 and 2007 in the course of HOS
rulemakings, constitute some of the most comprehensive sources of data
on driving hours, off-duty time, and utilization of the restart
provision. The scientific, operational, and economic analyses
underlying this rule have been meticulous and extensive. The provisions
made final today reflect both the paramount importance we attach to
safety and the critical role of the motor carrier industry in the U.S.
and world economy.
FMCSA's principal goal in the 2005 rule was to ensure truck drivers
had more opportunity for sleep than under the pre-2003 rule, and that
the typical work schedule would more nearly approximate the 24-hour
circadian ideal.\2\ Before 2003, drivers were required to take only 8
hours off duty before driving again. After leaving the terminal,
returning home, and taking care of personal or family matters, a driver
meeting the minimum requirements simply did not have enough time to get
the 7-8 hours of sleep needed to maintain alertness.\3\ By extending
the minimum off-duty period from 8 to 10 hours, as the 2003 and 2005
rules did, FMCSA ensured most drivers, even those operating on
compressed schedules, would be able to go home and deal with private
matters and still have sufficient time for a full sleep cycle. That
objective was preserved in the 2007 IFR and today's final rule.
---------------------------------------------------------------------------
\2\ See section F.2. Circadian Influences, in the 2005 rule, 70
FR 49992.
\3\ See section E.1. Sleep Loss/Restriction, 70 FR 49982-49983.
---------------------------------------------------------------------------
The 2005 rule also required drivers who use sleeper berths to take
8 consecutive hours in the berth and another 2 hours off duty or in the
berth, as the driver chose. The previous regulations allowed drivers to
split their sleeper-berth time into two periods, neither shorter than
two hours. The result was that sleeper-berth drivers often failed to
take a single, uninterrupted sleep period long enough to avert fatigue.
The higher crash rates reported for sleeper-berth drivers by the
National Transportation Safety Board reflected that fact.\4\ The 8-hour
sleeper-berth period adopted in 2005 meant that these drivers would be
subject for the first time to the same kind of rest requirements as all
other drivers. The DC Circuit upheld that portion of the 2005 rule in
its 2007 decision and it is accordingly unchanged by this final rule.
---------------------------------------------------------------------------
\4\ See section F.4. Split Sleep, 70 FR 49994.
---------------------------------------------------------------------------
To enhance the effect of increased off-duty time, the Agency also
reduced the driving window. Before 2003, the misnamed ``15-hour rule''
allowed driving within a 15-hour window after coming on duty--but off-
duty time taken during that work shift was not included in the 15
hours. The result was drivers possibly being at the wheel 18 or 20
hours after coming on duty, without having had any significant rest.\5\
The 2003 rule therefore allowed driving only within a fixed 14-hour
window after coming on duty; off-duty time no longer stopped the clock.
The combination of 10 hours off duty and a 14-hour driving window
greatly increased the number of drivers who would maintain something
close to a 24-hour schedule. Circadian regularity contributes to
fatigue-avoidance; \6\ the longer off-duty requirement and the shorter
driving window combined to improve significantly the likelihood that
truck drivers would be adequately rested before taking to the highway.
Sleeper-berth drivers, of course, are less likely to be on a 24-hour
cycle than other drivers, but the 2005 rule also improved their ability
to obtain adequate rest before driving.
---------------------------------------------------------------------------
\5\ ``A 1999 study of dry freight truckload carriers by the
Truckload Carriers Association (TCA) revealed that drivers spent
nearly seven hours waiting for each freight shipment that they
picked up and delivered'' [70 FR 49986]. Those hours were typically
excluded from the 15-hour ``limit.''
\6\ See discussion and sources cited in the 2005 rule, 70 FR
49992.
---------------------------------------------------------------------------
Although the 10-hour off-duty requirement and the 14-hour, non-
extendable driving window reduced the
[[Page 69569]]
risk of fatigue, these provisions simultaneously also imposed new
constraints on motor carriers. To offset these constraints while
ensuring the fatigue benefits are realized, the Agency determined it
could allow additional operational flexibility by permitting increased
driving hours without diminishing the safety benefits of the new
provisions. The 2003 rule therefore allowed driving no more than 11
hours following a period of at least 10 consecutive hours off duty,
compared to the 10 hours of driving permitted under the former
regulations. This balance reflects the integrated nature of the
Agency's approach to improving the HOS rules. The drivers who operated
instrumented vehicles in the large on-road study conducted by Hanowski
and his colleagues at the Virginia Tech Transportation Institute \7\
showed no statistically significant difference in ``critical
incidents'' between the 10th and 11th hour of driving. ``Critical
incidents'' were defined as crashes, near crashes (where a rapid
evasive maneuver was needed to avoid a crash) and crash-relevant
conflicts (which required a crash-avoidance maneuver less severe than a
near crash, but more severe than normal driving). This result strongly
supports the Agency's conclusion that an 11-hour driving limit adopted
in this final rule, when combined with the 14-hour driving window and
10 hours of off-duty time, does not pose an increased risk to safety.
---------------------------------------------------------------------------
\7\ See Hanowski et al., 2005, 2007a, and 2007b in the List of
References in section F.
---------------------------------------------------------------------------
Although the Agency did not expect a large percentage of drivers to
drive for 11 full hours, any more than they had previously driven for a
full 10 hours, the 11th driving hour would give drivers and carriers
affected by the new 14-hour window additional time to complete runs
that might once have been stretched out over 15 or more hours.
Subsequent comparison and analysis of field survey data collected by
FMCSA during compliance reviews in 2005 and 2007 has borne out these
expectations (see below). Although drivers and carriers are using the
11th driving hour more often than they did in 2005, a significant
majority of drivers represented in FMCSA's 2007 field survey (69
percent) still drove less than 10 hours during a typical daily shift.
In comments to the docket for the 2005 rule and the 2007 IFR, the great
majority of drivers and carriers who addressed this issue supported the
11-hour driving limit.
The most controversial element in the 2003 rule was the so-called
34-hour restart. The 2003 rule did not amend the long-standing 60- and
70-hour regulations, which provide that a driver may not drive after
being on duty more than 60 hours in 7 consecutive days, or in certain
cases, 70 hours in 8 consecutive days. However, it did allow drivers a
new method of complying with those limits. Under the 34-hour restart
provision, drivers can restart their calculations of the cumulative on-
duty time at any time after the driver has 34 consecutive hours off-
duty. However, because the restart provision was accompanied by an
increase from 8 to 10 consecutive hours of off-duty time, and a 14-
hour, non-extendable window within which drivers must complete all
driving during the work shift, the Agency concluded that the restart
would not decrease safety. Safety data published since 2003 has given
the Agency no cause to reconsider this conclusion.
Under the pre-2003 rule, drivers could operate under an 18-hour
``day'' by driving 10 hours, taking 8 hours off duty, and then
repeating that cycle. This schedule shifted the driver's sleep period
backward by 6 hours each day, disrupting the circadian rhythm and
interfering with the body's ability to sleep, even if adequate rest
hours were available. This compressed schedule also meant that drivers
would exhaust their 60 hours of on-duty time early on the fifth day and
have to wait almost 3 days until the sliding 7-day measuring period
again allowed driving. In fact, this extended rest period provided an
economic incentive for drivers to falsify their records of duty status
and begin driving again in defiance of the 60-hour rule.
The 2003 rule eliminated the 18-hour ``day.'' The most compressed
schedule allowed by the 2003 and 2005 rules is a 21-hour ``day,'' i.e.,
10 hours off duty followed by 11 hours of driving. This is
substantially more consistent with the normal circadian rhythm. A 21-
hour ``day'' also means that drivers cannot reach the 60- or 70-hour
limit as quickly as they could before 2003 and would not have to wait
as long at the end of the week before regaining compliance with the 60-
or 70-hour rule.
FMCSA examined the available research on the time needed to recover
from fatigue.\8\ As the Agency noted in the preamble to the 2005 rule,
the Transportation Research Board team, performing a literature search,
``found five studies that provided information regarding the recovery
time needed for CMV drivers after working a long week. Four of these
studies provide support for recovery periods of 34 hours or less, while
only one of these studies supports a recovery period longer than 34
hours. Two studies suggest that a single 24-hour period is sufficient
time for a driver to recover from any cumulative fatigue'' [70 FR
49994]. The length of a recovery period is not the only factor to
consider in assessing the adequacy of weekly work-rest schedules,
however. In preparing the 2005 rule, ``[t]he Agency attempted to
determine whether the added hours of recovery through the use of a 44-
hour recovery period, created a net benefit in reducing fatigue
compared to the potential negative impact on circadian rhythm of
establishing a rotating cycle. The Agency has determined that there is
no conclusive scientific data to guide it in determining which factor
(recovery time vs. circadian disruption) is more effective in
alleviating fatigue. In sum, in deciding to adopt a 34-hour recovery
period, the Agency considered that compliance with a 34-hour recovery
period results in a CMV driver restarting work at approximately the
same time of day as his or her prior shift. The 34-hour recovery period
also avoids the shifting of daytime to nighttime schedules, which
research indicates can disturb the circadian rhythm and decrease
alertness'' [70 FR 50024-50025, August 25, 2005]. A 34-hour period
gives a driver an opportunity for two consecutive 8-hour sleep periods
separated by a 16-hour period of wakefulness, plus an additional two
hours at some point in the 34-hour period. It allows circadian
regularity to be maintained and brings the driver back to work on
approximately the same schedule as before the restart. While there is
valid evidence that drivers who get 8 consecutive hours of sleep every
day should not develop cumulative fatigue at all,\9\ those who fail to
follow a regular sleep schedule will be able to ``zero out'' their
fatigue by taking 34 consecutive hours off duty.\10\
---------------------------------------------------------------------------
\8\ See the discussion and sources cited in the 2005 final rule,
70 FR 49994-49995, 50023-50026.
\9\ ``For most healthy adults an average of 7 to 8 hours of
sleep per 24-hour period has been shown to be sufficient to avoid
detrimental effects on performance.'' 70 FR 50016.
\10\ For example, a 1972 study ``involved subjects who worked 8
hours a day for 3 days, followed by a 4 hours on/4 hours off
schedule (similar to driving with a sleeper berth) over a 2-day
period. [The researcher] found that the average performance of
drivers dropped to 67 percent of baseline toward the end of this
period. A 24-hour rest period was sufficient to permit recovery back
to baseline. A simulator study examined daytime driving of 14 hours
on/10 hours off over a 15-day periods * * * These authors found that
24 hours was an adequate amount of time for recovery. A third study
* * * found a dramatic recovery with respect to fatigue in team
drivers who stopped overnight in the middle of a 4 to 5 day trip.
Thus, with less than 24 hours off, a single night of sleep was very
helpful for recovery.'' 70 FR 49994.
---------------------------------------------------------------------------
[[Page 69570]]
Opponents of the 34-hour restart argue that, if used to the maximum
over an extended period, it allows more driving and on-duty time on a
weekly basis than the pre-2003 rule. In theory this is true, but FMCSA
at that time concluded that the restart provision, like the 11th hour
of driving time, would not be utilized to the theoretical maximum
calculated by some commenters. Commenters have not provided nor has the
Agency seen any contrary evidence. As the Agency pointed out in the
preamble to the 2005 rule, use of the 34-hour restart to generate
routinely the very long driving and on-duty times critics fear--up to
84 hours on duty in 7 days or 98 hours in 8 days--requires an imaginary
world with ``nearly perfect logistics for picking up and delivering a
load * * * in other words, total elimination of waiting time to load,
mechanical and equipment problems, and traffic- and weather-related
delays.'' [70 FR 50022]. The Agency's more realistic scenarios have
been borne out by all recent evidence. As reported in the 2007 IFR [72
FR 71264-71265], data collected during FMCSA's 2007 field survey show
that use of the 11th driving hour by long-haul drivers increased only
slightly since the Agency's similar 2005 survey. In 2005, 23 percent of
the driving periods examined reached into the 11th hour, while 27
percent involved the 11th hour in 2007. Utilization of the restart
provision has also increased somewhat; 84 percent of the drivers in the
2007 sample \11\ took at least one restart period of 34 or more hours
during the typical work week, compared to 73 percent of drivers in the
comparable 2005 survey. Eight percent of the restart periods were
exactly 34 hours long in 2007, compared to 5 percent in 2005; 22
percent of the restart periods were between 36 and 44 hours long in
2007, exactly the same as in 2005; and 65 percent of the periods
exceeded 44 hours in 2007, compared to 68 percent in 2005. The Agency
concluded in its 2007 IFR that, ``while the restart provision is being
used by drivers, the average restart period is far longer than 34
hours'' [72 FR 71265].
---------------------------------------------------------------------------
\11\ A note printed below the table showing ``Restart Off-Duty
Hours'' on page 4 of the 2007 field survey incorrectly states that
``[t]his analysis excluded any restart period in excess of 72-
hours.'' In fact, the table shows that 22 percent of restart periods
were longer than 72 hours. That error does not affect comparisons
with restart periods from the 2005 field survey.
---------------------------------------------------------------------------
The American Trucking Associations (ATA) surveyed its members in
August, 2007, on their use of the 11th hour of driving time and the 34-
hour restart. The 69 carriers that responded represent 234,000 drivers;
46 percent of these drivers reported using the 11th driving hour
(meaning 54 percent did not), and driving into the 11th hour on 13
percent of their daily trips (meaning that 87 percent of the trips did
not reach the 11th hour). Because of the way ATA's restart data were
collected, they are difficult to compare to FMCSA data. Still, the
Agency concluded both in its 2007 IFR and in today's final rule that
the motor carrier industry is using both the 11th hour and the 34-hour
restart provision, not to maximize driving and on-duty time, but for
``operational flexibility,'' which is precisely its intended purpose
[Id.]. This conclusion is supported by data and comments submitted to
the docket, as well as data and analysis in the 2008 RIA accompanying
today's final rule. No commenters provided data or analysis indicating
any driver's actual use of the restart period over an extended period.
Operational data therefore provide no affirmation of concerns of
vastly increased on-duty and driving time as a consequence of the 11th
driving hour and the restart provision. The great majority of drivers
who addressed this issue in comments to the docket praised the 34-hour
restart provision and explained that it has enabled them to schedule
their work more flexibly than ever before and thus to spend more time
with their families, instead of waiting out the 60- or 70-hour clock at
some truck stop far from home. The 2007 FMCSA Field Survey continues to
support this conclusion. While miscellaneous off-duty periods taken by
drivers when confronting certain logistical realities cannot be used to
extend the 14-hour window within which up to 11 hours of driving time
may take place, these off-duty periods are not counted as driving time
or on-duty time and thus would reduce the likelihood of accumulating 84
hours on duty in 7 days, or 98 hours on-duty in 8 days. For example, if
a driver is delayed for a few hours while waiting to unload a shipment
and goes off duty, that off-duty time does not extend the 14-hour
window within which up to 11 hours of driving time may take place;
however, it is not counted in the maximum 60 or 70 hours of on-duty
time allowed within a 7- or 8-day consecutive period, or following a
minimum 34-hour restart period. In other words, it reduces the
likelihood that a driver would accumulate the maximum 84 hours on duty
in 7 days, or 98 hours on-duty in 8 days, as noted by commenters.
The preamble to the 2005 rule discussed in detail the various
effects that driving a CMV might have on the health of drivers (70 FR
49982-49992). There is some evidence of adverse effects; long-term
exposure to diesel exhaust seems to entail an increased risk of
cancer,\12\ and a recent study found an even higher incidence of
ischemic heart disease among unionized truck drivers.\13\ With the
exception of noise exposure, all of the studies related to driver
health and exposure lack a dose-response curve for the factor in
question that would allow FMCSA to estimate reliably the effect of
longer or shorter driving and on-duty time on driver health. As for
noise exposure, FMCSA concluded that drivers should not have
``significant hearing loss over a lifetime of on-the-job exposure, even
if drivers drove the maximum hours allowed by this final rule.'' \14\
---------------------------------------------------------------------------
\12\ Measured by standardized mortality ratios (SMRs) at a 95%
confidence interval. See next note.
\13\ Laden, et al., ``Cause-Specific Mortality in the Unionized
U.S. Trucking Industry,'' Environmental Health Perspectives, Vol.
115, No. 8, August 2007, pp. 1192-1196.
\14\ See section E.3., Exposure to Noise, in the 2005 final
rule, 70 FR 49987.
---------------------------------------------------------------------------
The Agency is required by statute to balance the benefits likely to
be achieved by its regulations against the costs likely to be
imposed.\15\ Because the likely benefits to driver health cannot be
estimated in the absence of a dose-response curve, the 2005 rule
declined to impose HOS limits designed to address health issues.
Although new health studies continue to be published, some directly
focused on truck drivers, the fundamental problem remains: the health
effects of any particular change in the HOS regulations are unknown.
---------------------------------------------------------------------------
\15\ 49 U.S.C. 31136(c)(2), 31502(d).
---------------------------------------------------------------------------
As explained in the Regulatory Impact Analysis (RIA) prepared for
the 2005 rule (and subsequent RIAs), the Agency examined several policy
scenarios (each of which included variations on the daily driving and
on-duty limits, minimum restart periods, and other options) and
developed elaborate simulation models of trucking operations, costs,
and fatigue-performance relationships [see 70 FR 50044 et seq.]. FMCSA
also undertook extensive analyses of the effect of various policy
scenarios on societal costs and benefits. This final rule, like the
2005 rule, adopts by far the most cost-beneficial of the policy
scenarios examined, as it is the only option that yielded net benefits.
For instance, the selected policy option yielded $270 million in
estimated net annual benefits versus net annual societal costs of
between $1 billion and $1.7 billion
[[Page 69571]]
annually for the other options considered.\16\ Analysis carried out
while preparing today's final rule estimated that eliminating the 11th
driving hour and 34-hour restart provisions would result in net annual
costs of $2.2 billion.\17\ The Agency also conducted sensitivity
analyses involving elimination of the 11th daily driving hour, both in
the 2005 and the 2008 final rules. In these analyses, the Agency
essentially doubled the likely percentage of fatigue-related large
truck crashes, tripled the value of a statistical life, and increased
by 40 percent the risk of a fatigue-related large truck crash in the
11th hour of driving; in all cases, however, the societal costs of
eliminating the 11th driving hour exceeded the benefits.\18\
---------------------------------------------------------------------------
\16\ See Section 6.7, exhibit 6-7, p. 71 of the 2005 HOS RIA.
\17\ See Executive Summary Exhibit ES-4, p. ES-6 of the 2008 HOS
RIA.
\18\ See Section 6.8, pp. 72-75 of the 2005 HOS RIA, and Section
6.7, pp. 67-71 of the 2008 HOS RIA.
---------------------------------------------------------------------------
This rulemaking rests on a wide-ranging body of data and
comprehensive analyses, and complies with all Congressional mandates.
By adopting HOS regulations that include increased daily off-duty time,
a shorter driving window, a longer period of uninterrupted rest for
sleeper-berth drivers, and sufficient time for two full sleep periods
before restarting the 60- or 70-hour clock, the rule ensures CMVs are
``operated safely'' and drivers' responsibilities ``do not impair their
ability to operate the vehicles safely,'' as required by 49 U.S.C.
31136(a)(1)-(2), respectively. FMCSA discussed in the preamble to the
2005 rule the possible ``deleterious effect [of driving] on the
physical condition of the operators'' (49 U.S.C. 31136(a)(4)),
concluding that any incremental health effects associated with the
additional driving and on-duty time theoretically available under the
rule cannot be calculated. This conclusion remains accurate today. The
RIA for this final rule fully complies with the statutory requirement
to consider benefits and costs of regulatory activities. Furthermore,
data on fatigue-related highway fatalities published since 2003 show
nominal annual fluctuations, but nothing of the rising trend implied by
some criticism of the IFR and related earlier rules. In fact, the
overall large truck fatality rate is at its lowest level ever.
D. Discussion of Comments
FMCSA received 913 comments to the docket; 26 submissions were
duplicates or non-germane letters. The majority of the commenters were
individuals, some of whom identified themselves as CMV drivers. In
addition, 122 motor carriers, 8 trucking associations, 17 other
industry trade associations, one State agency, and 6 safety advocacy
groups responded to the public docket. Over 300 comments expressed
general support for the IFR. Commenters who specifically addressed the
11-hour driving limit were divided, with 67 supporting it and 53
stating that 11 hours is too long. A number of commenters favored the
34-hour restart provision, with 164 indicating their support and 30
stating that it should be changed. As one carrier noted, ``The 34-hour
restart had an immediate impact on our drivers, not only did it allow
more flexibility, it improved their quality of life by providing them a
way to spend more time with their family, not only did it refresh their
available hours for working, but it refreshed them mentally and
physically which helps them be more alert and safer on the highways''
(FMCSA-2004-19608-3006). The American Moving and Storage Association
(AMSA) believed, ``The pro-safety aspects of the 34-hour recovery and
restart provision are undeniable.'' The ATA pointed out, ``Often a
restart is taken before weekly hours have expired in order to better
fit driving schedules to available work and driver preferences. This
element of the new rules is plainly doing more for stress reduction and
driver satisfaction than any other change.''
J.B. Hunt emphasized that the rule permits, but does not require,
drivers whose accumulated fatigue has been reduced or eliminated by the
34-hour restart to return to working, and it is a much safer system
than the previous rule. Also, if the driver has obtained restorative
rest eliminating accumulative fatigue, then it should not matter how
much more work the driver is able to do. J.B. Hunt noted that ``safety
reports tend to indicate that drivers are as safe or safer now than
they were before the current rules were put in place.'' The National
Industrial Transportation League (NITL) agreed, stating that, ``Many
shippers using private fleets have effectively built the 34-hour
restart into their operations, which has improved driver satisfaction
and quality of life.'' Two commenters argued that shorter (24-hour)
restart periods for oilfield and construction industry trucks under 49
CFR 395.1 have worked safely.
Two commenters, the National Private Truck Council (NPTC) and ATA,
stated that Federal data and anecdotal information showed that motor
carrier safety had improved since the 2003 rule was implemented. NPTC
stated that the Agency's interpretation of fatal accident data was
consistent with self-reported data it had submitted from 63 private
carriers. These companies had reported that their safety performance
improved in the first year of the rule. NPTC reported that the safety
records of these companies have remained steady in the subsequent
years. ATA cited safety improvements shown in national crash data and
mentioned that data from the Bureau of Labor Statistics (BLS) on truck
driver non-fatal occupational injuries show a large decline in truck
driver injuries from 2002 to 2006.
Over 90 carriers and carrier associations indicated that they had
seen no adverse safety impacts from the HOS regulations; in fact most
of them reported reductions in crashes and in injuries. After reviewing
its own safety statistics, Schneider National found ``significant
improvements in safety under the current HOS regulations. Safety is a
multi-variant challenge, and while we cannot attribute all of the
improvement to the HOS rules, it is reasonable to conclude that the
current HOS regulations have not had a negative impact on safety (as
some have suggested).'' The Canadian Trucking Alliance stated that
since the Canadian and U.S. restart provisions have been in place, they
have been widely used by Canadian drivers operating on both sides of
the border, and the carriers' safety experience has been positive. A
few carriers also noted that they have experienced no increase in
crashes during the 11th hour of driving. The Missouri DOT agreed that
the rules have not detrimentally impacted safety and that the 11-hour
limit and the 34-hour restart provisions should remain intact. One
driver believed that his company has had an increase in the number of
accidents because of the 11-hour driving limit, but provided no data to
support his assertion (FMCSA-2004-19608-3187).
ATA discussed truckload and less than truckload (LTL) carriers, and
noted that for truckload operations the 11th hour of driving provides
flexibility to dispatchers who work with customers and drivers to
schedule freight pickup and delivery times. More drivers are able to
take off-duty rest periods at home. The extra hour also provides a
cushion for irregular route drivers who deal with highway congestion,
weather, construction, and other unexpected delays. Some carrier
operations can provide dedicated service with fewer trucks and fewer
drivers than in the past. ATA pointed out that some LTL operations have
found opportunities to move freight more efficiently with a planned use
of the 11th hour. Over 60
[[Page 69572]]
carriers and carrier associations echoed these points. One carrier also
noted that the rule allows more effective management of equipment
maintenance (FMCSA-2004-19608-2878). The National Solid Waste
Management Association supported the rule provisions because its
members must deal with irregular collection schedules that include
Federal holidays. The Specialized Carriers and Rigging Association
asserted that its members must deal with coordinating various service
provider and customer requirements, and comply with government permit
restrictions on routes and operating times. Schneider National's
simulation modeling indicated that productivity would suffer if the 11-
hour limit and 34-hour restart were eliminated.
Many commenters raised issues not addressed in the IFR, which
focused only on the 11-hour driving limit and the 34-hour restart
provisions vacated by the Court. In particular, 212 commenters objected
to the non-extendable 14-hour duty period; and 259 expressed opposition
to the sleeper-berth provisions. Because FMCSA responded to these
concerns at length in the 2005 rule, and because the DC Circuit upheld
the Agency's sleeper-berth and ``14 hour rule'' provisions, the Agency
will not discuss them further in this document. See 70 FR 50012-50014,
50026-50031, August 25, 2005. Commenters also continued to express
concerns about shipper practices, parking, driver pay structure, and
other issues that are beyond the scope of the IFR and outside of
FMCSA's statutory authority. These comments are not addressed in this
final rule.
The following sections summarize the comments submitted in response
to the specific topics covered by the IFR. In particular, the
discussion addresses the comments of the Insurance Institute for
Highway Safety (IIHS), Public Citizen, and, in a joint filing,
Advocates for Highway and Auto Safety, Public Citizen, the
International Brotherhood of Teamsters, and the Truck Safety Coalition
(Advocates et al.).
1. Statutory Duty. Advocates et al. asserted that FMCSA has failed
to make safety its highest priority. They argued that ``The Agency has
relegated its safety mission to simply a balancing of economic costs to
industry without regard for its basic mission--to prevent deaths,
injuries, and adverse health impacts of much longer driving and working
hours.''
FMCSA Response
This rule is an excellent example of the paramount value FMCSA
places on safety. It significantly reduces the daily driving window and
lengthens the off-duty period drivers must take, greatly reducing the
risk of short- and long-term fatigue while providing operational
flexibility. However, as noted in the Legal Basis section, the Agency
must consider multiple factors in issuing any motor carrier regulation,
including their costs and benefits (49 U.S.C. 31136(c)(2)(A) and
31502(d)).
In its 2004 decision, the DC Circuit stated: ``* * * neither
petitioners nor the court suggests that the statute requires the agency
to protect driver health to the exclusion of [cost/benefit] factors,
only that the agency must consider it'' (374 F.3d at 1217). The Court
thus acknowledges that the Agency must consider costs, benefits, and
health factors in developing regulations, though it provided no further
guidance on the weight to assign each factor. There is no case law on
point, and the legislative history is silent. The FMCSA has therefore
used its analytical capacity, expertise, knowledge of the industry, and
best judgment to create a rule that enhances motor carrier safety while
minimizing costs, consistent with its primary safety mission.
2. Comments on Safety. Advocates et al., and to an extent some
private citizens and drivers, disputed the Agency's assertion that
motor carrier safety has improved since the 2003 rule went into effect,
and argued that national crash data contradict this claim.
Specifically, they stated that Fatality Analysis Reporting System
(FARS) data are not consistent and, at best, show insignificant
decreases in the fraction of fatigue-related crashes. In any case, they
noted that the preamble to the 2005 rule itself described recent crash
data as unreliable and inconclusive. They also pointed out that total
large truck crash fatalities only began to decline in 2006, and that
large truck fatal crashes per million vehicle miles traveled actually
increased immediately after the 2003 rule went into effect. Regardless
of national crash data, commenters stated that the Agency has not and
cannot establish any causal link between improved safety and the 34-
hour restart or the 11th hour of allowable driving time. Commenters
also criticized the Agency for not carefully studying the actual near-
term safety impact of the 2003 rule.
FMCSA Response
FMCSA has consistently been cautious about inferring causal
relationships between the HOS requirements and trends in overall motor
carrier safety. The Agency believes that the data show no decline in
highway safety since the implementation of the 2003 rule and its re-
adoption in the 2005 rule and the 2007 IFR.
The Agency also examined crashes per million vehicle miles
traveled. The fact that the fatal crash rate continues to follow a
downward trend suggests the HOS regulations have not had an adverse
impact on safety. The number of fatigue-related crashes is small and
has remained relatively stable from year to year, without any clear
trend since the 2003 rule was adopted.
Advocates et al. are inconsistent in arguing that national crash
data show a definite degradation in safety since the 2003 rule was
implemented, while simultaneously claiming that the data FMCSA has
cited are too preliminary and have not been studied in enough detail to
allow final conclusions. The Agency acknowledges that all data contain
``noise'' and that three or even four years' worth of annual crash data
may not provide definitive evidence of the effect of the current HOS
rule. Nonetheless, actual operations have thus far validated the
Agency's analysis of the benefits and costs of this rule.
FMCSA appreciates the self-reports of positive safety experiences
from carriers, discussed previously in this document and in the
preamble to the 2005 rule. While not definitive, these data are
consistent with the Agency's conclusion that safety has been maintained
under this rule. Moreover, the BLS data on occupational injuries
submitted by ATA seem to provide evidence of the overall improvement in
motor carrier safety. However, because these data are not linked in any
discernible manner to drivers' work-rest schedules, they do not provide
immediate justification for this rule.
3. Comments on Driver Health. Advocates et al. argued that FMCSA
has failed to protect driver health and consider the impact and costs
of long hours on driver health. They also stated that FMCSA's position
on driver health conflicted with the reports of the National Research
Council's Transportation Research Board (TRB) Panels of Experts. They
cited a newly published NIOSH Conference Report and Selective
Literature Review [Saltzman, G.M., and Belzer, M.H., (2007)]. Advocates
et al. also stated that there are decades of research on the impact of
long work hours or shift work which FMCSA failed to consider in its HOS
rule.
FMCSA Response
The Agency has considered driver health at length. FMCSA carefully
[[Page 69573]]
evaluated health impacts in the 2005 rule using three steps. First, the
Agency reviewed numerous studies, including those previously cited in a
May 20, 2000, notice of proposed rulemaking (NPRM) (65 FR 25540) and in
the 2003 rule. Second, FMCSA contracted with the TRB, which selected
nationally known health and fatigue experts to conduct a thorough
literature review of studies relevant to this rulemaking. The research
reports TRB discovered clustered around a few issues: Driver exposure
to diesel exhaust, noise, and vibration; prevalence of cardiovascular
disease; and the effect of sleep loss or deprivation, shift work, and
long work hours generally. As discussed at length in the preamble to
the 2005 rule (Section E, 70 FR 49982 et seq.), many of the studies
involved self-evaluations, which cannot be independently verified.
Other studies are based on objective data, but their results are not
``fine-grained,'' i.e., they do not allow the Agency to calculate the
health effects of a few more hours of driving or on-duty time, or a few
less. FMCSA reaffirms today the conclusion it reached in 2005--this
rule neither causes nor exacerbates the risks associated with driving a
CMV. Third, when commenters cited over 200 additional studies they
deemed relevant, TRB reviewed them to inform our health and safety
evaluation. In addition, the Agency conducted a literature review in
December of 2007 to review studies of driver fatigue and health that
were completed after the TRB review in 2005 [Belenky, G. and Wu, L.J.,
(2008)]. The Agency is not aware of, nor did any commenters provide,
any studies published since the 2005 rule that would change these
conclusions.
Advocates et al. appear to have misunderstood FMCSA's response to
the TRB panel. FMCSA did not dispute that there are some links between
driving and various health conditions. The TRB literature review on
driver health concluded that ``Lung cancer is likely caused by exposure
to diesel exhaust and the longer that exposure lasts the more likely it
is that a cancer will develop'' [Orris, P., et al. (2005), p. 8]. It
went on to state that while ``the evidence linking this exposure to
bladder cancer is less robust than that to lung cancer, it remains
likely that there is such a relationship and that it is governed by a
positive dose response curve'' (Id., p. 8). FMCSA has not disagreed
with this finding as explained fully in section E2, Exposure to Diesel
Exhaust, of the 2005 HOS rule. However, the Agency found: (1) That no
credible research exists which established a positive dose-response
curve between diesel exhaust and lung or bladder cancers; (2) that the
U.S. Environmental Protection Agency (EPA) has undertaken significant
steps to reduce the amount of diesel particulate matter to which
commercial drivers are exposed; and, (3) that no evidence of
significant increases in drivers' working hours has appeared, and
therefore, that drivers have experienced no increased exposure to
diesel exhaust as a result of the 2003 HOS rule compared to the prior
HOS rule. FMCSA concluded that, while exposure to diesel exhaust
probably entails some risk to drivers, neither the 2005 rule, the 2007
IFR, or this 2008 final rule causes or exacerbates that risk when
compared to prior HOS rules.
The TRB medical panel also concluded that there is some evidence
that cardiovascular disease (CVD) is caused in part by truck driving,
and its risk increases with the duration of this activity and the
disruption of the sleep cycle (Id., p. 8). In 2005, a NIOSH
representative to FMCSA's health group reviewed the literature
regarding CMV driving and the risk of developing CVD. The NIOSH
representative concluded, and FMCSA concurred, that current research
suggests the presence of only a weak association between CVD and truck
driving. Additionally, CVD is associated with many other occupational
types. No research studies were found that permitted an examination of
whether additional hours of driving a CMV impact driver health as
measured by increased CVD or acute myocardial infarction (AMI). FMCSA
therefore concluded that nothing in the available research implicates
today's HOS rule in a heightened risk of CVD or AMI.
The TRB medical panel concluded that based on exposure assessments,
noise-induced hearing loss could be a result of a working lifetime as a
driver (Id., p. 8). The Agency has previously funded research to test
the noise levels in large trucks and reviewed the documented research;
the tests and the research have not shown that truck noise exceeds OSHA
or FMCSA standards. The Agency is not aware of any data or
epidemiological evidence that the noise levels in CMVs may lead to
significant hearing loss.
Both the TRB medical panels and the Agency concluded that the
research on whole body vibration (WBV) and its potential health
effects, such as low back syndrome, is inconclusive because the studies
rely primarily on self-reporting and application of risks derived from
other environments (Id., p. 8). The literature related to commercial
driving and other musculoskeletal disorders suffers from the same
limitations (Id., p. 8). The studies that tested vibration in CMVs
found that vibration was close to the International Organization for
Standardization (ISO) health risk threshold, but it did not
consistently exceed that threshold. When comparing the 2003 HOS rule to
today's rule, it is the Agency's best judgment that, based on the
studies reviewed and comments received, WBV does not pose a significant
health risk to CMV drivers.
FMCSA also reviewed the NIOSH report entitled Overtime and Extended
Work Shifts: Recent Findings on Illnesses, Injuries, and Health
Behaviors [Caruso, C.C., et al. (2004)] and all studies in the report
regarding commercial drivers. For a complete description of the
Agency's review and analysis of the NIOSH report, see section E.6.,
Long Work Hours, in the preamble to the 2005 HOS rule (70 FR 49989 et
seq.). In short, the NIOSH review found that ``extended work shifts and
overtime lengthen exposure times and shorten recovery times, and the
health consequences are uncertain'' (Id., p. 29). The NIOSH review went
on to conclude that ``Despite the increased current interest in long
working hours, research questions remain about the ways overtime and
extended work shifts influence health and safety. Few studies have
examined how the number of hours worked per week, shift work, shift
length, the degree of control over one's work schedule, compensation
for overtime, and other characteristics of work schedules interact and
relate to health and safety'' (Id., p. 30). As a result of NIOSH's own
comments regarding the state of research, FMCSA concluded in 2005 and
again today that, based on current knowledge and the limited research
specific to CMV driver health and work hours, in the Agency's best
judgment there is not enough sufficient, credible evidence that the
number of work hours allowed by the HOS regulation will have a negative
impact on driver health.
Advocates et al. cited proceedings from a 2003 NIOSH conference
that were published in 2007 [Saltzman, G.M., and Belzer, M.H., (2007)].
FMCSA is well aware of the conference since FMCSA representatives
attended and presented papers at the meeting. The purpose of the
conference was to present research on driver health and to start a
dialogue on a National agenda for future research in the area of driver
health. Both anecdotal accounts and published research were included in
the NIOSH proceedings. The TRB literature
[[Page 69574]]
review of driver health included research published by January 2005.
Therefore, there is nothing new in these conference proceedings that
the Agency has not already considered.
In summary, as discussed at length in the 2005 rule, the Agency
undertook a comprehensive examination of issues related to driver
health. The Agency is aware of no new studies, nor have commenters
provided any, published since the 2005 rule was promulgated that have
changed these underlying conclusions and the regulatory provisions
adopted. Driver health research simply is not mature enough to allow
the conclusion that a number of extra hours of work would result in
increased driver health problems. Also, there are many confounding
factors that affect driver health, such as diet, smoking, and exercise.
It remains very difficult to isolate the impact of exposure and longer
working hours. The research to date has not provided a basis for
analyzing the health impact of the 2003 and 2005 final rules and the
2007 IFR, all of which allowed more driving time per day but fewer
hours of daily work and longer required off-duty periods. Without a
dose-response curve, which would indicate the incremental effect of
each hour of exposure to diesel exhaust, vibration or long working
hours, FMCSA has no basis for estimating health impacts and costs.
FMCSA, along with many other Federal and private entities, is funding
driver health research; however, it will be years before researchers
are able to separate the impacts of daily work exposure versus driver
lifestyle. The Agency concluded in 2005 that it was unable to quantify
or monetize the impacts of that rule on driver health; the same
conclusion applies to today's rule.
FMCSA also notes that several major carriers and associations,
including ATA, NPTC, and the National Industrial Transportation League
(NITL) stated that the rule had benefited drivers' health and quality
of life. These parties also noted that Advocates et al. focused
exclusively on two provisions of the rule, but ignored the changes
which provide drivers more time for sleep such as the 14-hour driving
window, the sleeper-berth rule, and especially the 10-hour off-duty
period. The 14-hour window limited the period of time available for
driving and, in combination with the 10-hour off-duty period, moved
drivers toward a 24-hour circadian period. Research at Virginia Tech
Transportation Institute (VTTI) has shown that as a result of the 2003
rule, CMV drivers are getting more sleep (1 hour) on a daily basis
[Hanowski, R.J., et al. (2005), p. 1]. Additionally, because the
sleeper-berth provision requires a consecutive 8-hour sleeper-berth
period and a second 2-hour off-duty or sleeper-berth period to be used
at the driver's discretion for breaks, naps, meals, and other personal
matters, drivers have a much greater opportunity to obtain additional
rest when needed. Consistent with the issues of exposure discussed
above, the Agency also was unable to quantify the positive impacts on
driver health from obtaining more sleep as a result of this rule.
Nonetheless, drivers are sleeping more overall with more circadian
regularity; and are now sleeping within normal ranges that are
consistent with a healthy lifestyle.
In an OOIDA survey, drivers reported an improved quality of life
based on the combined effects of the 2005 rule (70 FR 50025). A tally
of comments from the 2005 rule (Id. at 50037) leads to the same
conclusion. FMCSA agrees with comments emphasizing the need to treat
the rule as a single interactive whole, instead of analyzing its
provisions separately (70 FR 50041, 72 FR 71252). Moreover, numerous
drivers reported that the 2003 rule's off-duty time provided the
opportunity not only for sleep, but also for relaxation and personal
tasks that improved their quality of life (Id. at 50040). The preamble
to the 2005 rule also noted that certain lifestyle choices, over which
the Agency has no control, including eating, smoking, and exercise, may
``by themselves be predictive'' of cancer and cardiovascular disease.
(Id. at 50007). In addition, stress is a risk factor for cardiovascular
disease (Id. at 49988). To the extent today's rule can reduce stress,
it may be directly beneficial to driver health. The operational
flexibility allowed by the rule, which (to judge from their comments to
the docket) allows more drivers to spend weekends at home, may have
just such an effect.
4. Approach to Research. Advocates et al. stated that FMCSA
provided no evidence that it reviewed scientific research that did not
support its conclusions, and that the Agency disregarded almost all
studies not directly linked to truck driving or ignored studies on the
basis of flaws that were also evident in the few studies selected to
support the new HOS regulation.
FMCSA Response
FMCSA conducted extensive literature reviews in the course of
formulating its HOS regulations. In 1996, the then Office of Motor
Carriers (OMC) of the Federal Highway Administration (FHWA) collected,
reviewed, and docketed all relevant research on driver fatigue and
performance. As part of that effort, a detailed literature review on
driver fatigue was published (Freund, entry 956 in Docket 2350, Nov.
1999). In 1998, OMC commissioned an expert panel to deliberate on
changes to HOS regulations for commercial drivers. As discussed
previously, in 2005 FMCSA systematically and extensively researched
both U.S. and international health and fatigue studies and consulted
with other Federal safety and health experts. A detailed description of
the qualifications of the TRB team and the methodology used can be
found in the preamble to the 2005 rule (70 FR 49981). In addition, the
Agency conducted a literature review in December of 2007 of studies of
driver fatigue and health that were completed after the TRB review in
2005 [Belenky, G. and Wu, L.J., (2008)]. FMCSA has used the best
available research in its HOS rulemaking.
Comment. Advocates et al. stated that the FHWA Expert Panel did not
guide FMCSA decision-making on HOS. The Expert Panel was convened and
their report was published, but according to the commenters, the
Panel's findings were mostly disregarded or discarded, especially when
the findings contradicted the Agency's regulatory decisions.
FMCSA Response
The commenters are mistaken. The Expert Panel's role is to provide
guidance; it is exclusively the Agency's responsibility to make
decisions with regard to rulemaking. The Expert Panel is not
constrained by statutory requirements in undertaking its work, unlike
the Agency. Alternatively, the Agency must take into account various
statutory requirements in considering the guidance provided to it by
the Expert Panel, and make decisions based on this consideration. To
merely adopt recommendations by the Panel, without due consideration,
would be abdicating the Agency's statutory rulemaking responsibility.
Nonetheless, FMCSA did extensively use the Expert Panel's results to
guide its decision making.
The Expert Panel urged that a final rule rely on a 24-hour work/
rest cycle [Belenky, G., et al. (1998), p. 7]. This final rule is based
in part on that concept. A 14-hour driving window and a 10-hour off-
duty period, which are likely to be the standard for many drivers,
ensures 24-hour circadian regularity. The Expert Panel indicated that
``Off-duty hours must include enough continuous time off duty so that
drivers are able to meet the demands of life beyond their jobs and are
also able to obtain sufficient uninterrupted rest
[[Page 69575]]
* * * The time allotted for sleep must be a minimum of 9 hours'' (Id.,
p. 7). Today's final rule exceeds the Expert Panel's recommendation.
The Expert Panel noted that ``rest or sleep acquired in a sleeper
berth is not equivalent to rest or sleep in a bed'' (Id., p. 9). It
therefore urged that the Agency's regulations require an uninterrupted
sleeper-berth period of at least 7 hours to allow for 6 hours of
continuous sleep, with another period of at least 2 hours for a nap.
This was one of the many factors that went into the decision to change
the sleeper-berth provision in the 2005 rule. The 2005 rule required a
consecutive 8-hour sleeper-berth period to allow drivers to obtain one
primary period of sleep, and a second 2-hour off-duty or sleeper-berth
period to be used at the driver's discretion for breaks, naps, meals,
and other personal matters. This provision of the rule was upheld by
the DC Circuit and is therefore being adopted without change in today's
final rule.
The Expert Panel noted that ``recovery time periods must take into
consideration the necessity for overcoming cumulative fatigue resulting
from such schedules and must include sufficient sleep * * * Recovery
time should include at least two uninterrupted time periods * * * and
such recovery time must be made available at least once in every 7
days'' (Id., p. 12). The 2003 rule created a minimum 34-hour recovery
period that provides sufficient time for two 8-hour sleep periods and
one 16-hour period of intervening wakefulness, allowing the opportunity
for recovery from any potential cumulative fatigue that might occur.
Although the effect of the 34-hour restart cannot be isolated from all
the other factors that affect highway safety, it should be noted that
FMCSA's Field Surveys show increased use of the restart provision
between 2005 and 2007, at a time when the rate of fatigue-related fatal
truck crashes remained essentially unchanged and the overall large-
truck fatal crash rate dropped to the lowest level ever recorded. This
final rule therefore readopts the IFR's 34-hour restart provision
without change.
On one issue, the Expert Panel made a recommendation not fully
adopted by FMCSA; for example, the Panel stated that ``no distinction
should be made between on-duty time and driving time.'' The Panel noted
that ``for a variety of tasks (driving a bus, driving a truck,
operating a train), an early rise in accident risk * * * peaks between
2 and 4 hours after onset of duty, then falls and does not reach the
level of the early peak until after 12 to 14 hours, when it continues
to rise at an accelerating rate'' (Id., p. 8). This final rule, like
previous HOS rules, does distinguish between driving and on-duty time,
but today's 11-hour limit on driving time within a 14-hour on-duty
window is otherwise fully consistent with the Panel's conclusions.
The Agency did not reject the Expert Panel's recommendations; FMCSA
embraced the Expert Panel's report and developed a rule that is
supported by its recommendations
Comment. Advocates et al. also argued that FMCSA selectively quoted
from the studies it relied on to justify the HOS rule. Furthermore,
they stated that FMCSA has capriciously selected research studies,
relying on inadequate research and data to justify the IFR, while
rejecting conflicting studies.
FMCSA Response
On the contrary, FMCSA has worked on its current HOS rule for more
than a decade, and has funded considerable research to expand the
knowledge of sleep and fatigue science. This HOS rule has been
developed by FMCSA experts who have carefully reviewed and weighed the
findings from previous research efforts. Over the years the research
has improved as more sophisticated technology for data collection
became available. The Agency has relied and will continue to rely upon
improved research studies to produce the best possible regulations.
The first principle that the Agency uses in evaluating research is
that studies based on quantifiable, objective data that can be
independently verified and tested are preferable to those based on
subjective data such as individuals' opinions or perceptions. Where no
objective data that was collected through strictly controlled, unbiased
scientific experimentation exist, the Agency will use the best
alternatives available; that could, in some instances, be subjective
data. FMCSA prefers to use well-designed objective studies like the
Virginia Tech Transportation Institute (VTTI) naturalistic driving
research, rather than surveys of drivers.
The second principle is to rely primarily on independent studies
that are sufficient in scope, are peer reviewed, and use an application
of statistics (power analysis) to determine appropriate sample sizes.
The term ``sufficient in scope'' refers to the degree to which a study
is designed to answer the research questions posed, and the conclusions
can be reliably verified. Peer review is one of the important
procedures used to ensure that the quality of published information
meets the standards of the scientific and technical community. The
proper use of statistics ensures that results of a sample can be
generalized to a wider population.
The third principle used in evaluating research is to place greater
value on studies with repeatable findings or outcomes. Researchers test
theories, and the more these theories are validated, the more they are
generally accepted as principles. For example, the Driver Fatigue and
Alertness Study (DFAS) (Wylie et al., 1996) was the first to identify
the impact of circadian rhythm on CMV driver alertness, and almost
every fatigue study after the DFAS has used those results or found
similar results, to the point that the impact of circadian rhythm on
driver performance is now a generally accepted principle. This is
another reason that FMCSA relies on the findings by VTTI regarding TOT
versus hours of driving. Increasingly, naturalistic driving data and
studies are coming to the same conclusion--that time of day plays a
greater role in driver alertness than the number of hours driven. To
answer the concern of Advocates et al., FMCSA reviewed the research
literature on driver health and driver fatigue. The Agency used its
best judgment to weigh the adequacy of the research in developing the
2005 final rule. Because subsequent safety data have borne out that
judgment, today's rule adopts as final the provisions of the IFR.
Comment. Advocates et al. questioned the value of some of the
research that FMCSA used to justify its positions in the HOS rule.
Specifically, Advocates et al. stated that FMCSA relied on DFAS despite
critical comments from the Agency's peer review panel.
FMCSA Response
The commenter's argument that DFAS was deemed to have ``no
scientific credibility'' is belied by the fact that it formed the basis
for a paper published in the New England Journal of Medicine (Mitler,
et al., 1997).
DFAS was a landmark study of driver fatigue. Until recently, DFAS
was the largest on-road naturalistic driving study, with over 250,000
miles of driving data and over 80 CMV drivers operating in the United
States and Canada. It also was the first study that used sophisticated
technology to instrument trucks to measure driver performance and
fatigue. DFAS was particularly important in changing the methodology by
which commercial driver research would be conducted in the future,
introducing the use of instrumented vehicles and technology for
collecting data in a field setting. The DFAS findings also changed
commercial
[[Page 69576]]
driver fatigue research by identifying the important role of time of
day, rather than TOT, with regard to commercial driver fatigue.
In February 1995, a peer review of FHWA fatigue research was
performed by eleven specialists from the fields of human factors, sleep
research, behavioral psychology, and transportation safety. The peer
review actively examined several FHWA fatigue-related research projects
including DFAS. The peer review panel offered a number of comments on
how the study was conducted, and some of these could be construed as
criticisms. FHWA accepted some of these comments, but believed that
others reflected the individual research perspectives of the reviewers.
Many of the reviewers, particularly those whose primary experience was
in clinical or laboratory settings, were uncomfortable with the
operational aspects of the study. However, three other reviews of DFAS
by (primarily) applied researchers did not yield these kinds of
comments. As with all peer reviews, the DFAS authors and FHWA reviewed,
accepted, and made appropriate changes to the final report to reflect
constructive comments. Additionally, many of the reviewers commented
positively on the strength of the instrumentation package, the
extensive database obtained, and the ability of the team to obtain the
cooperation of the trucking industry. FHWA believed that the early and
intense scrutiny of DFAS findings by the peer review panel
significantly enhanced the long-term technical value of this project as
well as the planning of future fatigue research.
Comment. Advocates et al. stated that FMCSA also relied on the
findings of instrumented driving studies by Hanowski et al. at VTTI
(2005, 2007a, 2007b) to support the claim of no difference in risk
between the 11th and 10th hours of driving. Advocates were highly
critical of the study, stating that the drivers were using experimental
warning systems intended to alert them to signs of drowsiness, making
it impossible to isolate the effects of an additional hour of driving
or to generalize to the population of large truck drivers.
FMCSA Response
FMCSA referred in the IFR to a VTTI study (Hanowski et al., 2007b)
that was awaiting publication at the time. That study has now been peer
reviewed; the review is posted in the docket for this rule. The peer
review panel found that ``this study is a valuable contribution to the
science underpinning HOS regulations for truck drivers and has
potential application to other modes of transportation and operation as
well'' (Belenky et al., 2008). The peer review panel went on to say
that ``this is an excellent study and an excellent contribution to the
HOS debate'' (Id., p. 4). The statement from commenters that ``* * *
drivers were using experimental systems * * *'' is correct, but the
statement that ``[t]his makes it impossible to isolate the effects * *
*'' is not. Multiple analyses were conducted that did not include trips
where the experimental warning systems were used, thus enabling the
effects to be isolated for the additional hours of driving.
In the VTTI study, the number of trips or opportunities decreased
as the number of driving-hours increased. Drivers often concluded their
trips short of the 11th hour. However, there were 1,535 trips that did
include the 11th hour. All driving hours were treated the same. Drivers
who stopped a few minutes into the fourth, fifth, or eleventh hour were
classified as having driven in the fourth, fifth, or eleventh hour,
respectively. In other words, there was no ``systematic
underestimation'' of critical incident risk in the 11th hour of
driving; the same procedure was applied to all hours.
It is true that VTTI did not focus on drowsiness-related critical
incidents. By not excluding critical incidents based on their
underlying cause (for example, distraction or drowsiness), this study
directly addressed the driving-hour, or TOT, issue and allowed the
Agency to answer the question whether there is an increase in risk
associated with driving into the 11th hour. The multiple analyses that
parsed the data in many different ways consistently came to the same
conclusion: There is no measurable increased risk for drivers driving
in the 11th hour as compared to the 10th hour or any other driving-
hour. The finding that TOT is a poor predictor of crashes is consistent
with other well-conducted research in this domain (for example, Wylie
et al., 1996). FMCSA emphasizes that it relies on research like this
study to inform policy so that regulations are based on rigorous,
broadly-accepted, and repeatable protocols.
Comment. Advocates et al. stated that FMCSA contradicted itself by
rejecting the Paul Jovanis Final Report [Jovanis, P.P., et al. (2005)]
based on small sample size.
FMCSA Response
FMCSA contracted for and supported the Jovanis study. The Agency
did not use the Jovanis study in the 2005 HOS rulemaking because of
concerns with the models it used for crash risk in the later hours of
driving. For instance, there were 693 trips that involved the first
hour of driving and 28 crashes that occurred in the first hour of
driving (Id., p. 5). However, only 30 trips reached the 11th hour, with
four crashes occurring in the 11th hour of driving. Conversely, in the
VTTI study, there were 1,535 trips in the 11th hour of driving
[Hanowski. R.J., et al. (2007), p. 6]. As the result of this small
sample size in the 11th hour of driving, FMCSA believes that the models
produced in the Jovanis study lack predictive power about the crash
risk in the later hours of driving. In evaluating predictive power, one
need only inspect the standard error around the crash risk point-
estimate. The standard error increases with driving time, particularly
during hours 10 and 11. A model's standard error is the difference
between the actual values of the dependent variables (results) and the
predicted values. The range of the standard error suggests that the
confidence intervals around the model's point estimate could be similar
to the first hour of driving or as much as 11 times higher. The
confidence interval is used to indicate the reliability of the model's
point estimate. How likely the interval is to contain the parameter is
determined by the confidence level or confidence coefficient. This
model's very wide confidence interval in the 11th hour suggests that it
is not reliable for predicting driving risk in the 11th hour of
driving. Typically, logistical regression models with increasing
standard error or the funneling of confidence intervals around a point
estimate are indicative of sample size problems. The increasing
standard error demonstrates that this model has no real predictive
power in the later hours of driving. Therefore, FMCSA did not reject
the Jovanis study because it failed to support FMCSA's conclusions, but
rather because of technical problems with its underlying models.
Comment. Advocates et al. stated that Effects of Operating
Practices on Commercial Driver Alertness (O'Neil et al., 1999) suffers
from severe limitations and cannot be relied on by the Agency.
FMCSA Response
The O'Neil et al. study had appropriate and objective performance
metrics, and was tightly monitored and managed by the study agent. The
study consisted of laboratory research that assessed the effects of the
physical activity of loading and unloading on subsequent driver
alertness. This experiment also measured and documented drivers'
performance on a daily schedule involving 14 hours on
[[Page 69577]]
duty. The study concluded that ``drivers recovered to baseline
performance within 24 hours of the end of a driving week, and should be
fit to resume duty after 36 hours * * * [A] work schedule of 14 hours
on-duty/10 hours off-duty for a 5-day week did not appear to produce
cumulative fatigue. Subjective sleepiness, psychomotor vigilance
response, and some other measures showed a slight but statistically
significant deterioration over the 5-day driving week, but performance
on planned and unplanned driver challenge probes did not show
cumulative deterioration.'' [emphasis added]
Laboratory studies, like that of O'Neil and his colleagues, lack
the intense realism of naturalistic studies, but they compensate by
controlling variables more rigorously than is possible in over-the-road
operational studies. Both research strategies are valuable, and both
have limitations. FMCSA reviewed and took into account both types of
studies evaluated in this rulemaking and used its expert judgment in
deciding what weight to give any particular study.
Comment. Advocates et al. claimed that a 2006 study by the American
Transportation Research Institute (ATRI) has no credibility for HOS
regulatory decisions.
FMCSA Response
ATRI's ``Safety and Health Impacts of the New Hours-of-Service
Rules'' (Dick, V., et al., 2006) reviews how the 2005 HOS rule is
functioning within the trucking industry. The study examined aggregated
collision and driver injury data from motor carriers before and after
implementation of the 2003 HOS rule. The study was significant because
it involved 23 medium-to-large trucking fleets, roughly 100,000
commercial drivers and more than 10 billion vehicle miles of travel
each year. This study involves the largest number of drivers and
vehicle miles traveled that FMCSA is aware of. It found statistically
significant reductions in the overall collision rate per million
vehicle miles traveled (-3.7 percent), as well as reductions in the
preventable collision rate (-4.8 percent), the driver injury rate (-
12.6 percent), and the collision-related injury rate (-7.6 percent).
These results are consistent with the trends in the FARS data and
further support the conclusion that overall safety of the motor carrier
industry has been maintained since the 2003 and 2005 HOS rules became
effective.
Comment. Advocates et al. offered a 2007 study by Friswell and
Williamson that purported to show the dimensions of the fatigued truck
driver problem. The commenters asserted that this study contradicts
FMCSA's claim that fatigue contributes only to 7 percent of fatal large
truck crashes.
FMCSA Response
Friswell and Williamson (2007) found that 38 percent of CMV drivers
reported experiencing fatigue at least once a week. The authors
surveyed drivers who offered their opinions as to whether they had
experienced any fatigue during the prior week. The authors, however,
defined fatigue broadly ``as feeling drowsy, sleepy, tired, lethargic,
bored, unable to concentrate, unable to sustain attention, and mental
slowness.'' With this broad and subjective definition, it is impossible
to draw any reasonable scientific comparison with statistically-based
studies of fatigue relied upon by FMCSA, or to draw conclusions about
whether these drivers were so fatigued they could not safely drive a
CMV. Additionally, this study was of Australian drivers who operate
under very different operational and regulatory environments than in
the United States. Because of the vague definitions, subjective data,
and different operating and regulatory environments, the Agency did not
rely on the study for this rulemaking.
5. Use of Data, Analysis, and Modeling
a. Crash Data Used in the Regulatory Impact Analysis (RIA)
Advocates et al. criticized the Agency's use of Trucks Involved in
Fatal Accidents (TIFA) data in the 2007 IFR RIA. They argued that these
data have been used despite the Agency's acknowledgement of their
inherent limitations, because they show minimal increases in relative
risk between the 10th and 11th hours of driving. They also stated that
TIFA is primarily a survey, and depends critically on the ability of
the interviewers to retrieve police accident reports or to contact
someone directly connected with the fatal crash. Consequently, the TIFA
file does not parallel FARS in robustness.
Advocates et al. submitted a research paper by Gander et al. (2006)
on under-reporting of fatigue. This study, conducted from 2001 to 2002
in New Zealand, found that police correctly identified fatigue in only
41 to 71 percent of truck driver fatigue-related crashes. The Gander
study also emphasized that its findings of police report-based
underreporting of fatigue related truck crashes is generally accepted
throughout the research community across several countries.
FMCSA Response
FMCSA disagrees with the interpretation provided by Advocates et
al. of how the Agency has used data for its analyses. The purpose of
the RIA was to evaluate as comprehensively as possible the existing
data, models, and research findings to develop estimates of the costs
and benefits of implementing different policy options. The ultimate
goal was to inform Agency decision-making via the net benefits
associated with each option. Consequently, the Agency chose to use the
data it believed were best suited for this analysis. The TIFA data are
the most comprehensive source of data on fatal large truck crashes.
TIFA starts with FARS data and adds further information on those fatal
crashes. As FARS represents a census of all large truck fatal crashes,
TIFA also represents a census. As such, TIFA cannot reasonably be
characterized as less robust than FARS. It should also be noted that
the Agency did not restrict itself to FARS and TIFA data. As discussed
in section F (``Evaluation of Recent Safety and Operational Data Under
11-Hour and 34-Hour Rules'') of the 2007 IFR and sections H (Crash
Data) and I (Operational Data) of the 2005 Final Rule, FMCSA considered
a multitude of data sources in its deliberations.
FMCSA recognized in the RIA that accompanied the 2005 rule (pp. 44-
47) that TIFA and FARS data have inherent limitations. No dataset is
perfect, but at least FARS data, upon which TIFA is built, are
comprehensive. The datasets can be used to estimate costs and benefits
in a robust manner and, therefore, allowed the Agency to make informed
decisions about tradeoffs from different policy options. Non-linearity
of the crash risk over drive time is commonly found in all data of this
nature, and the Agency has never denied that the risk curve becomes
steeper at higher drive times in the analysis based on FARS and TIFA
data, even though the VTTI study showed no TOT effects. In fact, the
Agency developed two separate TOT models, the first for the 2005 rule,
and the second for the 2007 IFR, that are specifically designed to
track this feature in the data. Notwithstanding the dramatic effects
commenters infer from graphs of the data, empirically-based estimates
of the increase in crash risk in the 11th hour of drive time fail to
demonstrate that the safety benefits of eliminating that hour of drive
time exceed the costs.
The TIFA data are the only source of information on large truck
crashes by
[[Page 69578]]
hour of driving that also include coding of fatigue by an unbiased
source (the officer at the scene) or alternatively, data gathering and
validation by an independent researcher at the University of Michigan
Transportation Research Institute (UMTRI). FMCSA uses the national
level FARS/TIFA data precisely to minimize the potential impact of
State-by-State differences in the coding of fatigue by each State's
officers. There is no reason to believe that the coding of fatigue in
large truck crashes would change across all States simultaneously, such
that the national level estimates would vary significantly from one
year to the next. To ignore the FARS/TIFA data in the Agency's policy
making would be to omit from consideration the most reliable dataset
available on large truck fatal crashes.
Comment. Advocates et al. argued that the Agency mishandled FARS
and TIFA data and that the Agency did not explain its procedures and
rationale for dropping a large number of observations from the TIFA
dataset. They submitted a report by Quality Control Systems (QCS),
which criticized the Agency for excluding TIFA data with missing
fatigue information and suggested how to integrate records with partial
or missing information into the analysis.
FMCSA Response
The Agency does not agree with the suggestions in the QCS report
submitted by the commenters. First, the calculations performed on
FMCSA's dataset for the 2005, 2007, and 2008 RIAs, which use only
records with full HOS data, are not significantly different,
statistically speaking, from those that use the dataset that includes
records with partial HOS data, which QCS suggests to use. Second, the
standard errors from imputing missing information, whatever the
approach, must be integrated into the model. Given that the difference
between the partial and full data is not statistically significant, the
net result would be to degrade the performance of the model, not
enhance its precision.
b. Analysis and Modeling Used in the RIA
This section of the preamble outlines and responds to comments and
questions submitted to the docket regarding particular data, models,
and/or analysis used in the RIA.
i. Use of Default ``% Fatigue Crash'' Estimate. Advocates et al.
contended that the 1.6 percent of crashes attributed to fatigue that is
reported in the FARS data is unrealistic, and submitted numerous
criticisms and studies attempting to show that fatigue is
systematically underreported in police accident reports. They also
asserted that the Agency itself has acknowledged that the effect of
fatigue on crash risk is probably underestimated because it may often
play a less direct role in triggering a crash; fatigued individuals are
prone to a variety of mental and physical errors. These commenters
noted a systematic underreporting of fatigue that was even more evident
from the State-by-State FARS data, which showed that some States were
reporting relatively large numbers of crashes, but no fatigue.
Advocates et al. questioned the Agency's estimated baseline for
fatigue-related crashes (7 percent of all large truck crashes),
asserting that the choice was arbitrary.
FMCSA Response
FMCSA has already acknowledged that FARS tends to underreport
fatigue at a national level and that the coding of fatigue is a complex
determination based on a number of factors. Consequently, the current
RIA adjusts its estimate of the fatigue-related crash rate upward based
on Agency analysis of the available data.
As explained in the RIAs that accompanied the 2005 rule and the
2007 IFR, the Agency's 7 percent figure for fatigue-related truck
crashes within the long-haul sector was based on a series of
calculations using nationally representative data; the original
calculations were discussed in the RIA that accompanied the 2003 rule.
Specifically, the 7 percent baseline used in the 2005 and 2007 HOS RIAs
was calculated based on extensive analysis of FARS, General Estimates
System (GES), and other data to remove cases that would erroneously
bias the estimate downwards (for example, States that never coded
fatigue), and an increase to allow for inattention crashes likely to be
caused by fatigue. These calculations, assumptions, and analyses
associated with defining the baseline for fatigue-related large crashes
were explained in detail in Chapter 8 of the RIA that accompanied the
2003 rule, which is available in the docket. Overviews of these
calculations, with explanations of adjustments, were outlined in
Chapter 5 of both the 2005 and 2007 RIAs, and were also included in the
RIA accompanying today's final rule.
At a broader level, FMCSA is not aware of any studies that
unambiguously show what percentage of crashes are caused by fatigue.
Studies showing high levels of fatigue, such as the 40 percent figure
in a docketed National Transportation Safety Board (NTSB) study
(``Factors That Affect Fatigue in Heavy Truck Accidents: Volume 1
Analysis,'' January 18, 1995), were not derived from representative
samples of large truck crashes, but from highly atypical data sets. For
example, some of these studies are based primarily on fatal crashes in
which the truck driver was killed, which are more likely to involve
single-vehicle crashes, thereby increasing the percent of total crashes
likely to be fatigue-related. As such, these crashes are far from a
representative sample of all large truck crashes involving fatigue.
To address the uncertainty surrounding the baseline estimate, FMCSA
conducted sensitivity analyses in all of its RIAs in which it nearly
doubled the baseline estimate for fatigue-related crashes. However,
even these adjustments do not alter the conclusions of the RIA.
Additionally, while FMCSA had used an estimate of 7 percent as the
percentage of all large truck crashes across all hours of driving that
would be fatigue-related under the latest HOS rules, recent empirical
data actually indicate that this percentage may be relatively accurate.
Specifically, as is discussed in more depth later in this preamble,
recent TIFA data (i.e., calendar years 2004 through 2006, or after the
latest HOS rules became effective) reveal that the percentage of large
trucks involved in fatal crashes where the large truck driver was coded
as fatigued was only 2.2 percent (or one of 45 large trucks involved in
fatal crashes in the 11th driving hour between calendar years 2004 and
2006). Given the relative-risk curves estimated and used in the present
RIA analysis, one would naturally expect the percent of fatigue-related
involvements in the 11th hour of driving to be higher than 7 percent,
since the average across all driving hours was estimated at 7 percent.
However, since the percent derived from recent empirical data indicates
a much lower percent (2.2%), FMCSA analysts believe the original
analysis regarding the 7 percent figure is accurate, even when
recognizing that the coding of fatigue-related crashes may be
underestimated.
ii. Calculation of Relative Risk Ratios Used in RIA. Advocates et
al. expressed confusion over the use of TIFA data in deriving ratios to
estimate the relative risk of a fatigue-related truck crash by hour of
driving. In particular, they noted that the bar charts shown in Exhibit
5-1 of both the 2005 and 2007 RIAs were not consistent with the data
appearing above the bar charts in Exhibit 5-1, or with the text that
followed the Exhibit describing the calculation of relative risk ratios
used in the RIA.
[[Page 69579]]
Advocates et al. believed that the relative risk ratios calculated
by hour of driving identified along the X-axis of Exhibit 5-1 should
have been calculated by dividing by the relative risk associated with
the first hour of driving.
FMCSA Response
The bar charts in Exhibit 5-1 represent the relative risk ratios by
hour of driving, but the data appearing above those bar charts simply
represent the raw TIFA data which were used to calculate those ratios.
The text under the Exhibit describes the bar charts (representing the
relative risk ratios by driving hour) and not the raw data appearing
above the charts. This is explained via a step-by-step process
discussed on pp. 41-42 of the RIA for the 2007 IFR, which culminated in
the following statements on page 42:
Finally, to estimate the relative risk ratios that appear in
Exhibit 5-1, the percent of all trucks where fatigue was present at
the crash within each driving hour (i.e., 9.6 percent in the 11th
driving hour) was divided by 1.9 percent, or the percent of all
trucks involved in fatal crashes across all driving hours where it
was determined that the truck driver was fatigued at the crash. The
result is a relative risk estimate per involvement in a fatigue-
related crash for each driving hour. In the case of the 11th driving
hour, this estimate is equal to about five (or 9.6% divided by
1.9%), which is represented by the height of the bar chart in
Exhibit 5-1 for the 11th driving hour.
As explained in the preamble of the 2007 IFR and in the RIA
accompanying today's final rule, the appropriate baseline for
calculating the relative risk ratios for driving hours (most
importantly, the 11th driving hour) is not the first hour of driving,
but a combined weighted average of driving hours 1 through 10. If the
daily driving limit was restricted to 10 hours, thereby eliminating the
possibility of the 11th hour of daily driving, then that foregone 11th
hour would be redistributed to other drivers and/or other driving
shifts represented by the spectrum of allowable driving hours 1 through
10.
iii. Questions on the Use of the TOT Curve in 2005 Rule. Advocates
et al. expressed concern about the use of the TIFA data in calculating
the TOT curve for the purposes of estimating the safety benefits from
eliminating the 11th hour of daily driving. According to the
commenters, use of TIFA data underestimates the role of fatigue in
large truck crashes because TIFA is an extension of FARS data (that is,
it simply appends to FARS data additional information on the driver and
the large truck involved in the crash) and there have been numerous
concerns expressed about under-reporting of fatigue in FARs via use of
police accident reports. According to Advocates et al., this
subsequently led to very little difference in relative fatigue crash
risks between the 10th and 11th driving hours, while ignoring other
studies showing very high incidence of fatigue in crashes.
FMCSA Response
In previous HOS documents (the 2003 and 2005 RIAs, and the preamble
to the 2005 rule), and as discussed earlier in this document (`Use of
``% Fatigue Crash'' Estimates'), FMCSA acknowledged the potential
under-reporting of fatigue-related crashes in datasets such as FARS and
TIFA. Despite these limitations, FMCSA, as well as commenters to this
docket, have recognized the unique value of FARS and TIFA to motor
carrier policy makers. TIFA is the only dataset that the Agency is
aware of that represents a census of nationwide large truck fatal
crashes where fatigue is coded by hour of driving, a critical factor
for this rulemaking, which must consider differences in relative risk
of fatigue in large truck crashes between the 10th and 11th hour. No
other database available today provides such comprehensive information.
Additionally, an important element of the analysis was the relative
difference by hour of driving in the percent of large truck fatal
crashes where fatigue was present versus those where it was not, aside
from the baseline level of fatigue crashes. It is this relative
difference that forms the basis for calculating relative risk ratios,
calculating a TOT multiplier, and subsequently comparing marginal
differences in benefits and costs associated with setting a daily
driving limit at, for example 10 versus 11 hours. It is true that there
are relatively few data points in TIFA involving high-duration TOT such
that one could reliably estimate relative risks of fatigue-related
crashes. But there are enough total data points at both low and high
levels to find a reliable statistical relationship that FMCSA can use
to derive a curve for interpolating a relative risk value for the 11th
hour and then develop an associated TOT multiplier. This is precisely
what the Agency did in its 2005 and 2007 RIAs, as discussed extensively
in Section E (``Evaluation of Issues Concerning the Regulatory Impact
Analysis'') of its 2007 HOS IFR. Dr. M. Laurentius Marais, PhD, an
independent statistician asked by ATA to review the entire issue,
concluded that FMCSA's approach ``has a reasonable basis, in contrast
with [Public Citizen's] illustrative example, which is virtually
guaranteed to produce a biased result.'' Dr. Marais found that
``FMCSA's cubic regression curve matches the curves produced by more
sophisticated methods quite closely over the relevant range of driving
hours, in contrast to [Public Citizen's] illustrative alternative
curve, which departs substantially from the curves produced by more
sophisticated methods'' [72 FR 71254, fn. 2].
With regard to using other studies, such as Jovanis' work, to
develop relative risk ratios by hour of driving, it should be noted
that they have their own limitations, and reveal results that vary
widely. Jovanis (2005) presents an analysis of data from earlier years
that shows that all TOT levels beyond 4 hours are essentially the same,
and, if extrapolated, would give risks at the 11th hour of driving that
differed very little from those at 10 hours. In particular, this is the
case with the recent study by R.W. Hall and A. Mukherjee
(Transportation Research Part E44 (2008)) submitted to the docket by
commenters. The commenters stated that:
Driving-hour data from Lin et al. (1993; 1994), Park et al.
(2005), and Jovanis et al. (2005) are more likely to be accurate
than TIFA data because they were supplied by unionized carriers with
fixed routes and schedules who are less likely to have hours-of-
service violations. FMCSA's statistical models should have used
these and other strong studies of crash risk (e.g., Jones and Stein,
1987). One of these studies is by Hall and Mukherjee (2008,
attached), who conclude that the benefit of changing the driving
hour limit from 11 to 10 would be a 2 percent reduction in crashes.
A careful review of the paper by Hall and Mukherjee (who also
happened to be a co-author, along with Park, of the Jovanis Final
Report) reveals that the commenters appear to have misinterpreted the
paper's result. Their estimate of a 2-percent crash reduction from
reducing the daily driving limit to 10 hours was erroneously based on a
numerical example, not Hall and Mukherjee's actual estimate. For
instance:
We will use the normal distribution as an example, with
coefficient of variation values of 0.15 and 0.3, mean trip lengths
ranging from 2 to 8 h, and upper bounds ranging from 6 to 12 h. But
first, we use a mean driving time of 8 h and a standard deviation of
2.4 h (CV = 0.3) for illustration. [emphasis added] (Hall and
Muhkerjee 2008, p. 305)
Using these parameters for their numerical example, Hall and Mukherjee
find, in Table 2b of the paper, that the change from 11 hours down to
10 reduces crashes not by 1.99% (which is the reduction from 12 hours)
but by (1.99%-1.08%) or only 0.91%.
[[Page 69580]]
Furthermore, Hall and Mukherjee's quantitative estimate of the value of
reducing daily driving limits from 11 hours to 10 hours was, as shown
in Table 7 of the paper, $274 million minus $174 million, or $100
million per year, which is entirely consistent with FMCSA's estimates.
Hall and Mukherjee also stated that:
from an economic perspective, very stringent HOS rules, limiting
drivers to perhaps six hours per day, would reduce the cost of
crashes by no more than about $1.2 billion per year. This number is
consistent with prior FMCSA analyses, which estimated the annual
cost of fatigue-related crashes to be $2.3 billion per year. (p.
312)
One reason Hall and Mukherjee find relatively moderate benefits from
limiting daily driving hours is that they base their TOT function on
the study by Park et al. (Park, 2005) of less-than-truckload drivers.
That study did indeed find that crash risks in the 10th hour were twice
as high as in the first--but largely because the first hour was so far
below the average.
As seen in Table 3 of Jovanis (2005), beyond the first 4 hours, the
crash risks relative to the first hour are 1.865, 1.825, 1.837, 1.969,
1.741, and 2.108--all of these are significantly above the first hour,
but show no strong trend. The Jovanis Final Report states:
Importantly, the risk trend with driving time differs in
comparison to earlier findings (e.g. Lin, et al., 1993): the risk
increase after hour 4 * * * is not nearly as steep, particularly in
the last hour of driving. While unable to statistically
differentiate the crash risk, the trend in risk is a general
increase from hours 5 through 10. (Jovanis Final Report, p. 15)
Hall and Mukherjee extrapolate the trend from these data to the 11th
hour and beyond using the function y = 0.374 ln(x) +1.149 (from Hall
and Mukherjee, Figure 1); substituting 10.5 for x in this expression to
estimate risks during the 11th hour relative to the first hour yields
2.03.
The RIA's estimate of the value of crash reductions showed that
each one percent of long-haul sector crashes is worth $340 million.
Shifting 2 percent (or 0.02) of long-haul driving from the 11th hour of
driving of a trip to new trips lasting an average of 7 hours, would
reduce risks by about 0.35 \19\ times 0.02 which is 0.007, or 0.7
percent. In other words, the reduction in crash risk calculated using
this method would be only one-third as large as commenters claim--0.7
percent rather than 2 percent. If reducing long-haul sector crashes by
1 percent is worth $340 million, then reducing them by 0.7 percent
would be worth about $237 million (or $340 million times [0.7 percent
divided by 1 percent]). This is quite comparable to the value
calculated using FMCSA's method as discussed in the 2005 and 2007 RIAs.
This estimate relies on an extrapolation of Park's results, which
extended only through the 10th hour, as only the most recent Jovanis
study included the 11th hour explicitly. In that more recent study, the
uncertainty surrounding the 11th hour estimate was so great that it is
not distinguishable from the results based on the earlier study.
However one evaluates the Hall and Mukherjee paper, its conclusions
remain inconsistent with the results of the Virginia Tech
Transportation Institute \20\ studies which showed no difference in
crash precursors or critical incidents between the 10th and 11th hours
of driving.
---------------------------------------------------------------------------
\19\ This figure is derived by dividing 2.03 (Hall and
Mukherjee's relative risk in the 11th hour compared to the first
hour) by 1.505 (their average risk of the first seven hours).
\20\ See Hanowski et al., 2007a and 2007b in the List of
References in section F.
---------------------------------------------------------------------------
iv. Balkin Study. Advocates et al. also questioned FMCSA's reliance
on the Balkin study (Balkin, T., et al. (2000)) to estimate the
relationship between fatigue and performance, contending that the
study's ``primary finding was that there was no compensatory or
adaptive response by the subjects to even mild sleep loss, including
the 7-hour cohort.'' They also asserted that there were many
limitations to this study, including the way it was designed, so that
they could not calculate TOT effects for each sleep-restricted subject
hour-by-hour during the awake daytime period. Also, it was noted that
the sample size was small, the study failed to address circadian
rhythm, and finally that the sleep-restricted 7-hour cohort that
achieved an average of 6.28 hours of sleep did not regain baseline
performance even after three nights of sleep in a row.
FMCSA Response
The assertion that ``there was no compensatory or adaptive response
* * * to even mild sleep loss'' was not the primary finding of the
Balkin study. Rather than being ``centered on demonstrating recovery
related to the amount of sleep taken for each day,'' the Balkin report
states that ``the focus was on quantification of the relationship
between nighttime sleep duration and subsequent performance across 7
consecutive days,'' which was precisely the purpose for which FMCSA
used the results of this study--to develop a quantitative relationship
between sleep histories and alertness, for the purposes of predicting
changes in crash risks in a complex environment.
The Agency acknowledges that the Balkin study did not address TOT
issues. That is why FMCSA altered the basic approach of the Walter Reed
Sleep Performance Model (WRSPM), which was based on the Balkin study,
to use a separate TOT function/multiplier. The fact that the Balkin
study did not address TOT does not show that its approach to other
fatigue issues is incorrect and cannot be used in combination with a
TOT function.
Furthermore, contrary to commenters' assertion that the Balkin
study did not consider circadian rhythm, the study did indeed consider
circadian effects; data from the study (performance scores by time of
day) were used in calibrating the WRSPM, which uses circadian factors
as one component of predicted performance.
The Balkin study does state in its Executive Summary that
``following more severe sleep restriction (e.g., the 3-hour group)
recovery of performance was not complete after 3 consecutive nights of
recovery sleep (with 8 hours spent in bed on each night). This suggests
that full recovery from substantial sleep debt requires recovery sleep
of extended duration.'' [emphasis added.] The Balkin study clearly
indicates, however, that the 7-hour group is an example of a ``* * *
mild degree of sleep loss.'' The report's conclusions about recovery
from mild sleep loss are more equivocal than the observations about
severe or substantial sleep loss shown in the Executive Summary (see p.
2-85--``The effects of recovery sleep were variable * * * when
performance did recover, it was generally not complete after the first
8-hour recovery sleep period * * *. [I]n the 3-hour group, three 8-hour
recovery sleep periods were sometimes insufficient to restore
performance to baseline levels * * *'').
v. Appropriateness of the Sleep, Activity, Fatigue and Task
Effectiveness Model and the Fatigue Avoidance Scheduling Tool (SAFTE/
FAST Model). Advocates et al. questioned whether the SAFTE/FAST Model
used by FMCSA to estimate the impact of work schedules on fatigue and
subsequent performance levels under the latest HOS regulations was
appropriate for this analysis, given that it was originally designed as
a model used by the military. As such, there were questions as to
whether it met FMCSA's criteria that truck driving-related studies
should be used for this work. The commenters also pointed out that the
data integrated into the model represented a small sample size, which
FMCSA had considered disqualifying in
[[Page 69581]]
other contexts. It was noted that the SAFTE/FAST Model measures subject
recovery from sleep restriction in relation to a ``sleep reservoir'' on
which the subject draws to perform. Advocates et al. stated that the
average amount of sleep that hard-working drivers were asserted to get
per day, 6.57 hours, is inadequate, given FMCSA's repeated documented
belief that a minimum of 7 to 8 hours daily sleep is needed by most
individuals. It was also asserted that FMCSA ignored the caveat of the
Hursh et al. (Hursh 2004) article, which said of the SAFTE/FAST Model
that ``great care must be taken when applying a model to a performance
metric distinct from the one used to design the model.''
FMCSA Response
The SAFTE/FAST Model was based closely on the Balkin study, which
explicitly used truck drivers as subjects and assessed their
performance using truck driving simulators, not battle simulators.
Additionally, FMCSA worked directly with Dr. Steven Hursh, the
developer of the SAFTE/FAST Model, when it leased the model for use in
preparing its RIA. At that time, Dr. Hursh voiced no objections to the
use of SAFTE/FAST in the context of this work.
Dr. Greg Belenky, one of the authors of the Balkin study, stated in
a peer review of the 2007 HOS RIA (contained in the docket) that ``[i]t
makes excellent sense to embed the SAFTE/FAST model in a broader system
of fatigue risk assessment as is done here. * * * Conceptually using
the SAFTE/FAST model to evaluate schedules is a sound approach to
fatigue and crash risk assessment.'' Dr. Belenky obviously believed
that the SAFTE/FAST Model was an appropriate choice for modeling truck
drivers' responses to various schedules.
It is also important to clarify here that FMCSA did not claim to
use the SAFTE/FAST Model directly for TOT--rather, the results from
that model were augmented to increase the fatigue levels it predicted
by a factor derived from a separate data set (TIFA), via the TOT
multiplier exercise, in direct response to concerns raised by the DC
Circuit in 2004. The comment that the SAFTE/FAST Model was
inappropriate has no bearing on whether the TOT modeling was
reasonable. As with the Balkin study, SAFTE/FAST cannot account for the
effects of TOT. Again, that is why a separate TOT function was
developed and overlaid on the SAFTE/FAST results. The FMCSA is not
aware of any models that can simultaneously take into account all the
effects of sleep and schedule patterns, and the Agency therefore used
its expertise and best judgment to construct a method that would come
as close as possible to this ideal.
The SAFTE/FAST Model was used not to ``produce a new more palatable
crash risk analysis,'' but to incorporate the mechanism of the WRSPM
using a commercially available package that could reduce the chances of
error in application and also take into account shifts in the circadian
rhythms. The predicted sleep obtained during a series of 14-hour days
would not prevent a decline in performance from an ideal level; the
decline is small enough, however (given the chance to obtain recovery
sleep during a weekly break), to minimize the decrease in performance.
Lastly, the performance metric used was performance vigilant test (PVT)
scores, which the WRSPM showed were closely related to simulated
driving performance.
In summary, the SAFTE/FAST Model was based closely on a carefully
controlled laboratory study of the effects of important fatigue-related
factors on the performance and alertness of dozens of truck drivers. It
performed, overall, better than any other model tested against real-
world data. To this model, FMCSA added a TOT multiplier to further
improve its ability to assess the distinction between 10- and 11-hour
rule variants. Though Advocates et al. suggested other studies that
might be used to replace the particular TOT model that was used to
augment the SAFTE/FAST model results, they provided no indication as to
how the use of a different TOT function would solve the problem they
identify--that no TOT function has been integrated into a model that
can simultaneously account for all important schedule-related factors.
Asserting that FMCSA can estimate the effects of its policies only with
a model that has been developed to include all important factors, and
then has been empirically tested under real-world conditions (which
would require extremely intrusive monitoring of actual sleep and
performance, with a population large enough to produce a large sample
of serious crashes under enough combinations of schedules to
demonstrate that the model is accurate under all conditions) is to set
the bar unreasonably high.
vi. Applicability of the Cost/Benefit Analysis. Advocates et al.
stated that FMCSA had not assessed the costs and benefits for a
comprehensive set of commercial truck driver schedules under the new
HOS regulations, or more precisely, ``of truck drivers working all duty
hours both in a shift and over several consecutive days of driving and
other work.'' Conversely, ATA, in its filing to the docket, commented
that ``FMCSA has taken diligent and extraordinary steps to assure the
comprehensiveness of the [cost-benefit] analysis and its parts.'' NERA
Economic Consulting, as a result of its technical review of the 2007
IFR RIA, remarked in its docket comments that ``FMCSA has performed a
thorough, well-documented analysis of the costs and benefits of the
11th-hour and restart provisions. In fact, we have rarely seen such an
exhaustive and technically advanced analysis of a proposed rule from
any government agency.''
FMCSA Response
Contrary to the assertions of Advocates et al., FMCSA actually
evaluated the driving and working schedules of 12 distinct commercial
driver types representing all major industry operating segments (for
example, for-hire random trip drivers, private regular route drivers,
and team drivers) using a simulation model that maximized driver
productivity given certain pre-defined constraints (for example,
driving hour limits). The driver schedules were estimated over the
course of an entire year, so the model examined truck driver driving
and working hours over many days and many weeks and measured average
daily driving hours, average weekly working hours, and average restart
periods. Results from the simulation modeling were described in the
RIAs that accompanied the 2003 and 2005 rules and the 2007 IFR, and the
technical spreadsheet model and outputs were placed in the docket and
made available to the public by FMCSA. For additional details on these
models, see ``Section D. Regulatory Analysis and Notices'' later in
this notice and the stand-alone RIA for this final rule found in the
docket.
vii. Use of Outdated Crash Cost Estimates. Advocates et al.
criticized the Agency for use of crash cost estimates in the 2007 IFR
RIA that it claimed were not current.
FMCSA Response
In response to this comment, FMCSA incorporated several updates
into its 2008 HOS RIA to reflect more recent information that is
publicly available about crash costs and industry size. These updates
did not represent significant changes to the RIA and its findings, but
were made to reflect more recent information available on crash costs,
the value of a statistical life, and the size of the industry.
Specifically, these included the following:
[[Page 69582]]
Updating the estimates of the number of commercial drivers
engaged in long-haul operations;
Increasing the value of a statistical life, based on
updated values announced by the U.S. Department of Transportation in
February 2008; and
Updating the costs and benefit estimates of the RIA to
reflect 2005 dollars (from 2004 dollars) and incorporating new data on
crash damages.
More details on these and other nominal changes to the RIA are fully
explained in the RIA itself. None of these changes affected the cost/
benefit conclusions of the 2007 RIA.
c. Large Truck Crash Causation Study (LTCCS) Data and Analysis
In the 2007 IFR RIA, the Agency used data from the LTCCS. The
Agency received several comments about the data, many of them highly
technical. The Missouri Department of Transportation stated that the
LTCCS showed fatigue at the bottom of the top 10 causes of crashes. Dr.
Ronald R. Knipling stated in comments to the docket that the LTCCS
provided new, valid information on truck driver fatigue. Nevertheless,
he stated that single-vehicle crashes were over-sampled (or over-
weighted), which may have led to an exaggeration of truck driver
fatigue as a crash factor. According to the commenter, the data could
not be used to calculate true relative risk statistics because they
lacked exposure data. He also reiterated a critique (referred to as the
``confounding factors'' critique) that because driver schedule factors
were used in the fatigue determination, the fatigue variable and the
schedule factors were highly collinear, but the causal relationship was
less certain. Safety advocacy groups submitted numerous reviews from
the time of the study's inception and data gathering. These reviews
included both positive assessments and criticism, which the safety
groups highlighted. Deficiencies they cited included the lack of a
control group and exposure data; small sample size; missing, deficient,
and uncertain data; and lack of control for confounding factors. These
groups also repeated the criticisms of certain reviewers that the study
lacked a focused research design.
FMCSA Response
The LTCCS was designed to study why crashes that occurred did
occur. That is a characteristic of the data, not a flaw. As such, the
LTCCS is well-suited to investigate the causes of crashes or the
prevalence of contributory factors, both of which the Agency has
analyzed. Safety advocacy groups submitted several early reviews of the
LTCCS that were both favorable and unfavorable. Although some of the
reviewers faulted the study for the lack of a specific line of inquiry,
another reviewer correctly pointed out that there is nothing wrong with
studies with broad analytic objectives; almost all major economic
surveys conducted by government agencies collect broad amounts of data
to support several different lines of analysis. Nevertheless, as one
LTCCS reviewer points out, the study is focused insofar as it is
designed to gather information on the factors affecting large truck
crashes. The fact that the study was not designed to answer a specific
question about crash causation does not invalidate its use, although it
cannot be used to investigate every hypothesis with the same level of
accuracy.
As Knipling pointed out in his comment to the docket, the LTCCS
does make a meaningful contribution to research on driver fatigue, and
the Agency believes it does have useful applications for specific
analyses on this topic. The LTCCS sample size was not small; for a
study of this kind, approximately 1,000 crashes involving over 1,200
truck drivers is a very substantial sample. Certain lines of inquiry
may be limited by missing data or infrequent occurrences in the data,
but in the case of driver fatigue, the dataset yields 706 observations
(individual truck drivers) from 642 crashes. The safety advocacy groups
misunderstood the ``confounding factors'' critique. The LTCCS assessed
fatigue based on, among other things, drive time, not on an independent
physiological determination. This, coupled with Knipling's observation
that the apparent over-representation of single vehicle crashes would
exaggerate the role of fatigue, suggests that the study would
overstate, not understate, the importance of fatigue because an
investigator might have a bias toward coding fatigue where a driver has
long drive times, measures of work hours, etc., while the converse
might be true for low drive times and measures of work hours.
d. Supplemental LTCCS Analysis. The Agency recently commissioned a
study of the effects of fatigue in crashes included in the LTCCS. This
research has been placed in the docket for this rule, and a summary of
the analysis and results follows.
The LTCCS collected data on a random sample of approximately 1,000
crashes involving at least one large truck (gross vehicle weight rating
of at least 10,000 pounds) during 2001-2003 where there was a fatality,
an incapacitating injury, or a non-incapacitating, but evident injury.
The study was a nationwide survey with 24 data collection sites in 17
states and the results were weighted to represent all nationwide
crashes. For each crash, investigators collected data on all vehicles
involved, including information from driver, witness, and police
interviews and from driver logbooks, and determined the critical reason
for the crash. Critical reason is not an assignment of fault, but an
assessment of whether driver behavior fostered the occurrence of the
crash. The LTCCS also provides information on the driver's level of
attention, behavior, and mental or emotional state prior to the crash,
including an assessment of fatigue. Investigators determined whether
each driver was fatigued based on the driver interview and other
information such as logbooks. Factors such as fatigue may have been
present even if the driver had not been assigned any critical reason
for the crash. Even though a driver may have been found to be fatigued,
he or she may not have had any responsibility for the crash.
This analysis focused on the truck drivers involved in the crashes.
FMCSA used logistic regression to investigate the relationship between
driver fatigue and driver-related critical reason and several
explanatory variables: hours of driving, hours worked on day of crash,
hours awake, hours of last sleep, hours worked last week, time of day,
number of vehicles involved, day of week, and truck type. Because not
all fatigued drivers were assigned the critical reason for a crash, the
analysis of critical reason more directly examines how the explanatory
variables cause crashes. Hours of driving and hours worked provide
insight into TOT effects, while hours worked last week can determine
the extent to which cumulative fatigue exists. The most important
variables associated with driver fatigue were hours awake, hours of
last sleep, hours worked the previous week, and the number of vehicles
involved. The most important variables associated with driver critical
reason were hours of last sleep, hours worked last week, number of
vehicles involved, and truck type.
This analysis revealed several interesting facts. Among the more
striking findings are that sleep-related variables (including time
awake, length of last sleep, and average sleep over the past week) are
clearly related to both the chance that a driver of a large truck
involved in a crash was fatigued and to the chance that the driver was
assigned the critical reason for the crash. (See
[[Page 69583]]
figures 3, 4, and 16 in the analysis for plots of the chance of fatigue
against these three variables, respectively, and figures 9, 10, and 17
for plots of the chance that the driver was critically responsible
against the same three variables.) At the same time, though driving
extra long hours in a day or working overtime the previous week
appeared to increase fatigue (see figures 1 and 5 for plots of the
chance of fatigue against these two variables, respectively), there was
no evidence that they increased the chance that a driver was assigned
the critical reason for the crash (see figures 7 and 11 of the analysis
for plots of the chance that the driver was critically responsible
against the same two variables); that estimated probability was almost
constant at the longer hours. Furthermore, the main model seemed to
validate an hypothesis of the peer review panel for the RIA of this
rule that time awake and total on-duty time were more critical than
driving time. Long hours of driving in a day did not appear to be
related even to fatigue, once hours awake and hours worked were taken
into account (see figure 6).
e. New Fatal Large Truck Crash Data. Section F (``Evaluation of
Recent Safety and Operational Data Under 11-Hour and 34-Hour Rules'')
of the 2007 IFR provided an extensive discussion of FARS data
considered in this rulemaking. In that discussion, FMCSA included a
table showing FARS fatal, and fatigue-related fatal, large truck crash
data for calendar years 2000 through 2006. Additionally, FMCSA stated
that, ``In the 3 years since the 2003 HOS rule has been in effect, the
number of fatigue-related large truck crashes as a percent of all large
truck fatal crashes each year has remained relatively stable,''
fluctuating ``from a high of 2.2 percent in 2000 to a low of 1.5
percent in 2001 and 2004.'' Since the issuance of the 2007 IFR, the
National Highway Traffic Safety Administration has released its 2007
FARS Annual File via the FARS Web site (http://www-fars.nhtsa.dot.gov/Main/index.aspx). Those data indicate that the percent of fatal large
truck crashes has continued to remain relatively stable, with 78 fatal
crashes where the driver of the large truck was coded as fatigued, out
of a total of 4,190 large truck fatal crashes in 2007. Thus, 1.9
percent of all fatal large truck crashes occurring in 2007 involved a
fatigued truck driver, well within the longer-term high and low of 2.2
percent in calendar year 2000 and 1.5 percent in calendar year 2004.
Section F of the 2007 IFR also included a discussion of Trucks
Involved in Fatal Accidents (TIFA) data for the calendar years 1991
through 2005. As described in the IFR, the TIFA data file combines
large truck fatal crash data obtained annually from NHTSA's FARS with
other information obtained by the University of Michigan Transportation
Research Institute (UMTRI). UMTRI collects the additional data items on
the commercial driver and vehicle through telephone interviews with
truck drivers, carriers, or investigating officers after fatal crashes.
UMTRI combines vehicle, crash, and occupant records from FARS with
information obtained through TIFA, such as the physical configuration
of the large truck, the motor carrier's operating authority, and the
hour of daily driving at the time of the crash.
TIFA and FARS variables of particular interest include whether the
large truck driver was coded as fatigued at the time of the crash, the
time of day, the intended trip distance, and hours driving since the
last mandatory off-duty period (a legal minimum of 8 hours for data
through calendar year 2003 and 10 hours for calendar year 2004 and 2005
data).
TIFA data published in the 2007 IFR covered the years 1991 through
2005 (the most recent data then available). This file represents more
than 50,000 medium/heavy trucks involved in fatal crashes in the U.S.;
the truck driver was fatigued in approximately 1,000 of these crashes.
The TIFA data covering calendar year 2006 have become available
this year and show a continued downward trend in fatigue-related fatal
crashes since the Agency published the 2003 HOS rule (see Table 1).
Table 1--Large Trucks Involved in Fatal and Fatigue-Related Fatal
Crashes in the 11th Hour of Driving, by Calendar Year
------------------------------------------------------------------------
Fatigue-
coded Fatigue-
Calendar year (CY) Fatal (large coded as
crashes truck percent of
driver) total
------------------------------------------------------------------------
1991-2002........................ 94 9 9.6
2003............................. 13 1 7.7
2004............................. 16 0 0.0
2005............................. 13 1 7.7
2006............................. 16 0 0.0
1991-2003 Combined............... 107 10 9.3
2004-2006 Combined............... 45 1 2.2
------------------------------------------------------------------------
Source: Trucks Involved in Fatal Accidents (TIFA), 1991-2006.
In CY 2006, 16 large trucks were involved in fatal crashes where
the driver was operating in the 11th hour, but in none of these cases
was the truck driver coded as fatigued. Combining the 2004-2006 TIFA
indicates that there were a total of 45 large trucks involved in fatal
crashes during the 11th hour of driving, of which one of these (or 2.2
percent) involved a truck driver coded as fatigued. Conversely,
combining data for the 1991-2003 period indicates there were a total of
107 large trucks involved in fatal crashes during the 11th hour of
driving, of which 10 involved a fatigued truck driver. As such,
collectively, the 2004 through 2006 TIFA data represent a significant
improvement over the pre-2003 period, in terms of the percentage of
large truck drivers operating in the 11th hour who were coded as
fatigued at the time of the fatal crash. Although only three years of
TIFA data are available since implementation of the new HOS rules at
the start of 2004, the trend is encouraging.
E. Regulatory Analyses and Notices
Executive Order 12866
FMCSA has determined that this action is an economically
significant regulatory action within the meaning of Executive Order
12866. As such, the Agency has prepared an RIA analyzing the costs and
benefits of this final rule. A copy of the RIA is included in the
docket referenced at the beginning of this final rule. However, a brief
summary of the RIA results is provided
[[Page 69584]]
in this section. The Office of Management and Budget (OMB) has reviewed
this document.
Introduction to the RIA
This analysis considers and assesses the potential consequences of
two potential regulatory options:
Option 1 is the current rule. It allows up to 11 hours of driving,
allows a new 7- or 8-day period to begin after a 34-hour restart break,
and some splitting of off-duty periods using sleeper berths. The option
constrains the use of sleeper berths, however, to ensure that the main
sleeper berth period is at least 8 hours long, supplemented by an
additional 2-hour break that may be taken outside the sleeper berth.
Option 2 is more stringent than Option 1, limiting driving to 10
(rather than 11) hours in a tour of duty, and eliminating the 34-hour
restart provision. The sleeper-berth provisions are the same as in
Option 1, and both options retain the short-haul provision contained in
the 2005 rule. That provision allows operators of short-haul vehicles
that do not require a commercial driver's license (CDL)--typically
those of less than 26,000 lbs gross vehicle weight rating (GVWR)--and
that remain within a 150 mile radius of their base, to keep timecards
in lieu of logbooks and be on-duty up to 16 consecutive hours two days
during a seven-day work week.
Overview of the RIA Analysis
The simulation model used to estimate the costs for implementing
Options 1 and 2 was first loaded with data representative of shipping
patterns and carrier cost structures, and tested to ensure that it
could realistically simulate typical lengths of haul, empty mile
ratios, and productivity. The model was then set up to cover the most
important cases, under certain constraints (that is, daily driving hour
limits, minimum restart hours) represented under each option, and used
to simulate carrier operations under different conditions and HOS rule
options. FMCSA then analyzed the data representing the simulated
operations, using changes in miles driven as a measure of productivity
impacts. Output measures from individual runs were weighted to give a
realistic representation of the affected industry, including the
drivers' use of the most important provisions of the options. The
weighted changes in productivity from this procedure were then used to
estimate the cost increases imposed on the industry by each option,
using an analysis of the changes in wages and other costs likely to
result from changes in productivity. These productivity-related costs
were combined with transition costs associated with shifting to new
regulations to produce estimates of total social costs.
Safety impacts were measured by feeding the on-duty and driving
schedules from the carrier simulation model into an operator fatigue
model (known as the SAFTE/FAST Model) to project driver performance
levels under different schedules allowed under each HOS option. Then,
the fatigue model results were used to estimate the resulting changes
in crash risks under each HOS option and for the different operations
cases. Changes in fatigue-related crash risks, calibrated to match
realistic levels, were then multiplied by the value of all affected
crashes to yield estimates of total benefits.
The approaches used to estimate the costs and benefits of this
final rule have not changed since the 2007 HOS IFR. However, several
inputs to the RIA were updated between the IFR and this final rule in
order to reflect the most recent data available. Specifically, these
updates include the following: dollar values are now expressed in 2005
dollars rather than 2004 dollars; the industry population has been
updated to account for growth in numbers of long-haul drivers over the
past six years, when the data were originally collected for the 2003
rule; estimated changes in productivity and crashes have been corrected
slightly to include effects on the less-than-truckload sector; and the
value of crash reductions has been updated using newer crash
information and a revised value of a statistical life. These updates
were made either because of comments submitted to the docket regarding
the outdated inputs used in the cost/benefit estimation, or, in the
case of a higher value of a statistical life, due to new guidance
issued by the DOT in Spring 2008.
RIA Results
The weighted productivity impacts from implementing Option 2 (that
is, 10 hours driving, no restart) results in a 7.30 percent reduction
in driver (labor) productivity compared to the current IFR. From
research conducted for the 2003 Rule RIA (contained in the docket),
FMCSA analysts showed that each one percent change in driver
productivity is associated with just under $300 million in costs using
a population estimate based on the year 2000 and cost figures expressed
in 2004 dollars. Updating to a more recent and larger 2005 estimate for
the long-haul driver population and expressed in terms of 2005 dollars
raises the cost of each one percent change in productivity to $335
million. Multiplying the weighted average productivity impacts by the
costs per percent decrease in productivity yields $2,443 million in
annual costs associated with implementing Option 2 (relative to Option
1, which obviously showed no change in costs relative to the current
operating environment).
The reduction in crash risk from implementing Option 2 instead of
Option 1 was estimated to be approximately 0.63 percent. This change in
risk was valued by multiplying it by an estimate of the total annual
damage associated with heavy-duty long-haul truck crashes, updated to
account for a slight increase in total crashes, and re-estimated
damages per crash using a higher value of a statistical life. This
total was multiplied by the percentage of total damages that were
caused by the long-haul segment, yielding just over $34 billion. The
reduction in risk attributable to Option 2, given this total value, is
about 0.63 percent x $34 billion or about $214 million per year.
In summary, the total annual costs from implementing Option 2 are
roughly $2,443 million and the total annual safety benefits are roughly
$214 million, resulting in a net annual cost from implementing Option 2
of approximately $2,229 million (in 2005 dollars).
The Agency conducted a series of sensitivity analyses, where it
``stress tested'' various assumptions related to elimination of the
11th hour of driving. Specifically, the Agency revised its assumptions
with regard to several important inputs to the RIA, including the
percent of all large truck crashes that are fatigue related (increasing
it from 7 percent to 15 percent), the value of a statistical life
(increasing it from $5.5 million to more than $10 million), and raising
the relative risk of a fatigue-related crash in the 11th hour of
driving (by 1.3 times the value used in the revised TOT multiplier).
Each change improved the safety benefits relative to costs from
eliminating the 11th hour of daily driving, but none of these changes
in individual assumptions made elimination of the 11th driving hour
cost beneficial. Although it is unlikely that FMCSA mis-specified these
three assumptions in its initial analysis, the Agency nonetheless
combined all of the new assumptions in a way that makes elimination of
the 11th daily driving hour more favorable from a benefit-cost analysis
perspective. This exercise still generated net annual costs of $71
million, meaning that eliminating the 11th hour is unlikely to be cost-
effective under any reasonable set of
[[Page 69585]]
circumstances. This represents a brief summary of the contents of the
RIA accompanying this final rule. Readers are encouraged to review the
full contents of the stand-alone 2008 HOS RIA contained in the docket
to this rulemaking.
Regulatory Flexibility Act
Under the Regulatory Flexibility Act (RFA), as amended by the Small
Business Regulatory Enforcement Fairness Act of 1996 (Pub. L. 104-121,
110 Stat. 857), FMCSA is not required to prepare a final regulatory
flexibility analysis under 5 U.S.C. 604(a) for this final rule because
the Agency has not issued a notice of proposed rulemaking prior to this
action. However, FMCSA believes the RFA impacts of this final rule were
adequately described by the 2005 rule; there are no changes here.
Unfunded Mandates Reform Act of 1995
This final rule will not impose an unfunded Federal mandate, as
defined by the Unfunded Mandates Reform Act of 1995 (2 U.S.C. 1532, et.
seq.), that will result in the expenditure by State, local, and tribal
governments, in the aggregate, or by the private sector, of $128.1
million or more in any one year.
Paperwork Reduction Act
This final rule does not alter the existing information collection
requests for HOS recordkeeping.
Environmental Justice
FMCSA evaluated the environmental effects of this final rule in
accordance with Executive Order 12898 and determined that there are no
environmental justice issues associated with its provisions or any
collective environmental impact resulting from its promulgation.
Environmental justice issues would be raised if there were
``disproportionate'' and ``high and adverse impact'' on minority or
low-income populations. None of the alternatives analyzed in the
Agency's environmental assessment, discussed under National
Environmental Policy Act, would result in high and adverse
environmental impacts.
National Environmental Policy Act
FMCSA prepared an environmental assessment (EA) of the IFR in
accordance with the National Environmental Policy Act of 1969 (NEPA)
(42 U.S.C. 4321, et seq., as amended), the FMCSA's NEPA Implementing
Procedures and Policy for Considering Environmental Impacts (FMCSA
Order 5610.1),\21\ the Council on Environmental Quality Regulations
(CEQ) regulations implementing NEPA (40 CFR parts 1500-1508), the DOT
Order 5610.C (September 18, 1979, as amended on July 13, 1982 and July
30, 1985), entitled ``Procedures for Considering Environmental
Impacts,'' and other pertinent environmental regulations, Executive
Orders, statutes, and laws for consideration of environmental impacts
of FMCSA actions. The Agency relies on all of the authorities noted in
this paragraph to ensure that it actively incorporates environmental
considerations into informed decision-making on all of its actions,
including rulemaking.
---------------------------------------------------------------------------
\21\ FMCSA's environmental procedures were published on March 1,
2004 (69 FR 9680), FMCSA Order 5610.1, National Environmental Policy
Act Implementing Procedures and Policy for Considering Environmental
Impacts, and effective on March 30, 2004.
---------------------------------------------------------------------------
As shown in the EA that accompanied the IFR, none of the
alternatives considered would have had a significant adverse impact on
the human environment. Subsequently, FMCSA determined that the IFR and
this final rule will not significantly affect the quality of the human
environment and that a comprehensive Environmental Impact Statement is
not required. The EA for the IFR, as well as the Agency's finding of no
significant impact (FONSI), are contained in the docket referenced at
the beginning of this notice.
Executive Order 13132 (Federalism)
This action has been analyzed in accordance with the principles and
criteria contained in Executive Order 13132. FMCSA has determined this
rule does not have a substantial direct effect on States, nor would it
limit the policymaking discretion of the States. Nothing in this
document preempts any State law or regulation.
Executive Order 12372 (Intergovernmental Review)
The regulations implementing Executive Order 12372 regarding
intergovernmental consultation on Federal programs and activities do
not apply to this program.
Executive Order 12630 (Taking of Private Property)
This rule will not effect a taking of private property or otherwise
have taking implications under Executive Order 12630, Governmental
Actions and Interference with Constitutionally Protected Property
Rights.
Executive Order 12988
This regulation meets the applicable standards set forth in
sections 3(a) and 3(b)(2) of Executive Order 12988, Civil Justice
Reform.
F. List of References
American Transportation Research Institute (ATRI), ``Safety and
Health Impacts of the New Hours-of-Service Rules,'' 2006.
Balkin, A.T., Thome, D., Sing, H., Thomas, M., Redmond, D.,
``Title Effects of Sleep Schedules on Commercial Motor Vehicle
Driver Performance'', Walter Reed Army Institute of Research, May
2000.
Belenky, G., McKnight, A.J., Mitler, M.M., Smiley, A., Tijerina,
L., Waller, P.F., Wierwille, W.W., and Willis, D.K., ``Potential
Hours-of-Service Regulations for Commercial Drivers: Report of the
Expert Panel,'' 1998.
Belenky, G., and Wu, L.J., ``Literature Review on Fatigue and
Health Issues Associated with Commercial Motor Vehicle Driver Hours
of Service: Update from 2004,'' Washington State University, 2008.
Caruso, C.C., Hitchcock, E.M., Dick, R.B., Russo, J.M., and
Schmit, J.M., ``Overtime and Extended Work Shifts: Recent Findings
on Illnesses, Injuries, and Health Behaviors,'' April 2004.
Dick, V., Knipling, R., and Hendrix, J, ``Safety and Health
Impacts of the New Hours-of-Service Rules,'' American Transportation
Research Institute (ATRI), March 2006.
Friswell, R., and Williamson, A., ``Exploratory study of fatigue
in light and short haul transport drivers in NSW, Australia,''
Accident Analysis Prevention. January 2008; 40(1):410-7. Aug 2007.
Hall, R.W. and Mukherjee, A., ``Bounds on effectiveness of
driver hours-of-service regulations for freight motor carriers,''
Elsevier Ltd. Transportation Research Part E 44, 2008, pp. 298-312.
Hanowski, R.J., Dingus, T.A., Sudweeks, J.D., Olson, R.L., and
Fumero, M.C., ``Assessment of the Revised Hours-of-Service
Regulations: Comparison of the 10th and 11th Hour of Driving Using
Critical Incident Data and Measuring Sleep Quantity Using Actigraphy
Data,'' Virginia Tech Transportation Institute, June 2005.
Hanowski, R.J., Olson, R.L., Bocanegra, J., Hickman, J.S.,
Dingus, T.A., and Sudweeks, J.D. ``Critical incidents that occur in
the 10th and 11th hour of driving in commercial vehicle operations:
Does risk increase in the 11th hour?,'' Arlington, VA: American
Transportation Research Institute. November, 2007.
Hanowski R.J., Olson R.L., Bocanegra, J., and Hickman, J.S.,
``Analysis of Risk as a Function of Driving-Hour: Assessment of
Driving Hours 1 Through 11'' Virginia Tech Transportation Institute,
December 2007.
Hursh, S.R., Redmond, D.P., Johnson, M.L., Thorne, D.R.,
Belenky, G., Balkin, T.L., Storm, W.F., Miller, J.C., and Eddy,
D.R., ``Fatigue models for Applied Research in Warfighting'',
Aviation Space and Environmental Medicine, Vol. 75, Number 3,
Supplement 2004.
Jovanis, P.P., Park, S-W., Gross, F., and Chen, K-Y., ``Crash
Risk and Hours Driving: Interim Report, FMCSA,'' Pennsylvania
Transportation Institute, Penn State University, February 2005.
[[Page 69586]]
Laden, Francine; Hart, Jaime E.; Smith, Thomas J.; Davis, Mary
E.; and Garshick, Eric. ``Cause-Specific Mortality in the Unionized
U.S. Trucking Industry,'' Environmental Health Perspectives, Vol.
115, No. 8, August, 2007, pp. 1192-1196.
O'Neil, T.R., Krueger, G.P., Van Hemel, S.B., and McGowan, A.L.,
``Effects of Operating Practices on Commercial Driver Alertness,''
1999.
Orris, P., Buchanan, D., Smiley, A., Davis, D., Dinges, D., and
Bergoffen, G., ``Literature Review on Health and Fatigue Issues
Associated with Commercial Motor Vehicle Driver Hours of Work,''
Transportation Research Board, 2005
Park, S., Mukherjee, A., Gross, F., Jovanis, P.P., ``Safety
Implications of Multi-day Driving Schedules for Truck Drivers:
Comparison of Field Experiments and Crash Data Analysis,'' 2005.
Saltzman G.M., and Belzer, M.H., ``Truck Driver Occupational
Safety and Health: 2003 Conference Report and Selective Literature
Review,'' February 2007.
Wylie, C.D., Shultz, T., Miller, J.C., Mitler, M.M., and Mackie,
R.R., ``Commercial Motor Vehicle Driver Fatigue and Alertness Study
(DFAS),'' 1996.
G. Removal of Rescission Provision
In view of the events following the 2003 HOS rule--a legal
challenge resulting in an adverse decision by the DC Circuit, which
Congress temporarily suspended to allow time for further Agency action,
culminating in a new rule--FMCSA thought it appropriate to highlight
that the pre-2003 rule had been entirely superseded. It did so by
promulgating Sec. 395.0 in the 2005 rule, which provided that ``[a]ny
regulations on hours of service of drivers in effect before April 28,
2003, which were amended or replaced by the final rule adopted on April
28, 2003 [69 FR 22456] are rescinded and not in effect.'' As there is
no longer a question that the pre-2003 rule is superseded, and further
absent any amendment of specific provisions of the Code of Federal
Regulations, the provision may not be consistent with the Federal
Register Act. FMCSA is therefore removing Sec. 395.0.
List of Subjects
49 CFR Part 385
Administrative practice and procedure, Highway safety, Motor
carriers, Motor vehicle safety, Reporting and recordkeeping
requirements.
49 CFR Part 395
Highway safety, Motor carriers, Reporting and recordkeeping
requirements.
0
In consideration of the foregoing, FMCSA adopts as final the interim
final rule published at 72 FR 71247, December 17, 2007, with the
following change:
PART 395--HOURS OF SERVICE OF DRIVERS
0
1. The authority citation for part 395 continues to read as follows:
Authority: 49 U.S.C. 504, 14122, 31133, 31136, 31502; Sec. 229,
Pub. L. 106-159, 113 Stat. 1748; Sec. 113, Pub. L. 103-311, 108
Stat. 1673, 1676; and 49 CFR 1.73.
Sec. 395.0 [Removed]
0
2. Remove Sec. 395.0.
Issued on: November 13, 2008.
David H. Hugel,
Deputy Administrator.
[FR Doc. E8-27437 Filed 11-18-08; 8:45 am]
BILLING CODE 4910-EX-P