Book Review

Traffic Safety

By Leonard Evans. (Pp 445; US$99.50) Bloomfield
Hills, MI, Science Serving Society, 2004. ISBN 0-
9754871-0-8.

Leonard Evans is a British-trained physicist who
retired after a career with General Motors and
became the self-appointed president of an entity
he calls Science Serving Society, the publisher of
this book. In other words, the book is self-
published. There is no indication of peer review
of the book manuscript, which might have
resulted in more objective and scientifically
sound work. He does mention presentations of
some of the material and has published some of
it in peer reviewed journals but there are gaping
chasms in his coverage of the science of road
injury reduction and unsound logic in some of his
analyses and prescriptions for amelioration of
injuries. Those desiring a comprehensive review
of what we know to enhance road safety will not
find it here.

Evans sensibly discusses mobility and safety,
risk homeostasis, error in belt use in police
records that bias belt effectiveness estimates
and irresponsible entertainment media. But there
are odd lapses, particularly for a physicist, such
as referral of high center of gravity (cg) as a
factor in rollover without mentioning track width,
i.e., distance between the center of the tires (t).
Higher cg only lowers vehicle stability if t is too
narrow. The risk increases as t/(2 x cg) is
lowered below 1.2, common for many pickup
trucks and utility vehicles that have high rollover
death rates. Evans claims that rollovers are
mainly a result of speeding but my study of speed
of vehicles in relation to t/2h found no correlation
at rollover sites (AJPH, 87:839-841, 1997). On the
test track, the lowest stability vehicles tip up in
90 degree turns at 25 miles per hour, sometimes
less. The dampened reduction in US death rates,
discussed below, is partly due to increased
rollover from unstable pickup trucks and utility
vehicles, the sales of which have increased
dramatically during the past two decades in the
US.

The main thesis of the book is that behavior
causes motor vehicle crashes and behavior must
be changed to reduce motor vehicle crashes.
Reducing incidence and severity of injury by
changing vehicles and environments is largely
dismissed as ineffective or not worth the costs.
Most of the epidemiological research on road
injuries and the effectiveness of
countermeasures is ignored or mentioned with
derision. Evans claims that evaluations of overall
effect of vehicle regulations overstate the effect
compared to specific studies of some of the of
crashworthiness and brake standards. He ignores
most of the crash prevention standards, such as
side running lights and reduced light reflection in
drivers' eyes, and considers only vehicle
occupants. The data, disaggregated by model
year and calendar year and corrected for vehicle
age, indicate reduced fatal collisions with
pedestrians and bicyclists as well. He claims
that such multi variate analysis is subject to
manipulation to obtain results desired by the
analyst. See my 1998 book, Injury Epidemiology,
for a detailed discussion of the studies of
regulation.

The double-paired comparison method that Evans
prefers is a complicated version of a case-control
study. It can be misused if choice of controls is
biased. For many analyses matching controls is
not possible. Most of his most contentious claims
are based on mere eyeballing of trends in injury
and death rates with no adjustment for factors
that he knows to be important such as
demographic shifts, population and vehicle
density in geographic areas and changing mix of
vehicles, including non motorized vehicles. A
more comprehensive discussion of the
importance of these factors can be found in the
recent report from WHO on reducing road
injuries. It's free -- click on the link.

In his emphasis on crash prevention, Evans does
not acknowledge that the large majority of
crashes in the US do not occur in geographic
areas where the large majority of severe injuries
occur. He recognizes that severe injuries occur
mainly in areas of low traffic density but not that
the large majority of crashes occur in high traffic
density areas. If we target our efforts on
preventing crashes without accounting for
severity, prevention of injury will be reduced
substantially. His major prescription for crash
prevention is widespread use of video cameras to
record violations automatically and identify
license tag numbers of illegal speeders and those
who ignore stop signs and red lights. Tickets
would necessarily be issued to the vehicles
owners who may not be the violators. There are
limited studies that indicate some effectiveness
of such devices in crash reduction at or near
where they are used but Evans says that use of
the devices engenders political resistence on the
grounds that they are mere revenue producers.
Evans says this resistence to their use could be
offset by returning the fines collected to all
motorists in the form of a check of about $10 per
year but there is no evidence to support such an
assertion. Nothing is said about whether the
devices should be deployed more where most
crashes occur or where most severe injuries
occur. Nothing is said about how many
intersections or other sections of roads should
have the devices for maximum benefit. He claims
that this approach would be more cost-beneficial
than other approaches but the only cost-benefit
analysis in the book is devoted to frontal air
bags, which he says are not cost beneficial. Only
costs to the economy are used in that analysis,
ignoring the non economic burden of injuries.
Nothing is said about side curtain air bags.

In a diatribe against lawyers, Evans claims that
the "air bag mandate" in the US was imposed by
Joan Claybrook (lawyer) in the Carter
Administration, egged on by Ralph Nader
(lawyer). In fact the only air bag mandate was
introduced by non lawyers Douglas Toms, head of
NHTSA and a former state motor vehicle
administrator and John Volpe, Secretary of
Transportation and a former owner of a
construction company, in the Nixon
administration. Because their mandate was not a
performance standard as required by US law, the
standard was later revised to require minimum
forces on crash dummies in frontal impacts at 30
miles per hour, which is not an "air bag
mandate". US automakers chose to use air bags
to meet the standard but were not compelled to
do so if they had chosen to design the vehicles
differently.

Evans repeatedly emphasizes injuries imposed by
air bags but says nothing about seat belt related
injuries such as decapitations by belts and
injuries to necks and spinal cords (the "seat belt
syndrome"). He does not recognize that if all
public health measures that have any adverse
effects on a small proportion the population were
abandoned, there would be few, if any, in use. He
dismisses the benefit of a major shift in
transportation of children to back seats in the US
that resulted from child injuries in front seats
from air bag deployments. He writes, "The child
is denied the better view from the front seat --
and may consequently grow up knowing less
about driving, with possible adverse effects on
safety."

Evans examines the declines in fatal injury rates
in the Australia, Canada, the UK and the US with
no adjustments for factors noted above, and
claims that the US is no longer has the lowest
fatal injury rates because of overemphasis on air
bags and other aspects of vehicle
crashworthiness. Close examination of some of
his graphs suggests that other factors were
operative. For example, the UK rate declined to
near the US rate in the early 1960s prior to
vehicle regulation in the US. The two rates
declined similarly during 1965-1991 and the UK
rate declined more in the 1990s. With the
exception of the frontal crash standard that was
rescinded by the Reagan Administration but
reversed by the federal courts, which resulted in
air bags, there has been little new vehicle
regulation in the US since the early 1980s. Evans
presents no budgets or other evidence of
differential efforts to influence driver behavior in
the countries he compares.

In his claim that the US neglected driver
behavior, Evans ignores increased US federal
grants to the states for highway safety during the
years of vehicle regulation. These were
efficacious in some instances and adverse in
another -- namely high school driver education.
During 1975-1979, the federal government gave
the states $39.2 million for high school driver
education and $602.6 million for other programs
(J. Pub. Heal. Policy, 5:376-386). Although the
driver education funds were subsequently cut in
response to research showing the adverse effect,
the other programs continue to be funded to this
day. Scholars in US academic injury prevention
centers and other academic safety departments
will be surprised at Evan''s claim that the "US
lacks academic safety institutions" and what
exists is "nearly all in biomechanics and
crashworthiness".

Much of US road safety policy is determined by
state governments and is not influenced by
vehicle regulation. One interesting comparison
that Evans could have considered is comparison
among US states. For example, Massachusetts
has one of the highest crash rates among the
states but had the lowest road fatality rate per
registered vehicle in 2002. The Massachusetts
rate was 26 percent, 49 percent and 38 percent
below the UK, Canadian and Australian rates,
respectively, as reported by Evans for that year.
Does that mean that Massachusetts has a
superior road safety program to other US states
(and the countries that Evan's claims have
superior programs)? I don't think any serious
student of injury epidemiology would make such
an inference without extensive research on
specific programs, corrected for other factors
such as age distribution of licensed drivers,
vehicle density, mix of vehicles in use and the
like.

The mission for injury control researchers is not
to argue whether behavior change vs. vehicle and
environmental change is superior nor to claim
that a given country''s injury control efforts,
based on gross rates, is superior to another''s. It
is to find behavioral, vehicle and environmental
approaches that are demonstrated scientifically
to reduce injury and are acceptable or can be
made acceptable within the culture of a given
society. This book fails that test. It is based on
precious little good science and does no service
for society.

Leon S. Robertson

Retired injury epidemiologist, Yale University

1071 W. Mountain Nugget Drive

Green Valley, AZ 85614

nanlee252000@yahoo.com