Crash Protection for Child Passengers
U M T R I
• University Of Michigan Transportation Research Institute • January–March 2012 • Volume 43, Number 1 •
Crash Protection for
Child Passengers:
Rationale for Best Practice
Child Restraint Use
and Effectiveness
Principles of
Restraint Systems
Child Restraint
Child Restraint
Child Boosters
and Belts
Child Restraint
U M T R I
• University Of Michigan Transportation Research Institute • January–March 2012 • Volume 43, Number 1 •
ISSN 0739 7100
UMTRI Director: Peter F. Sweatman
UMTRI Associate Directors: Jean T. Shope, Lawrence W. Schneider
UMTRI Business Administrator: Cathy Seay-Ostrowski
UMTRI Communications and Marketing Director: Francine Romine
UMTRI Library Head: Bob Sweet
Writer and Editor: Joyce Daniels
Cover Photography: Courtesy of Miriam Manary
Layout & Printing: U-M Printing Services
The UMTRI Research Review is published four times a year by the Research
Information and Publications Center of the University of Michigan Transportation
Research Institute, 2901 Baxter Road, Ann Arbor, Michigan 48109-2150
( To subscribe, see the subscription form on
the inside back cover. For change of address or deletion, please send us your
address label or email
Follow UMTRI
Editor’s note: This special issue of the UMTRI Research Review provides the latest
research findings on crash protection for child passengers. This issue updates the highly
regarded July-September 2000 issue (volume 31, number 3) devoted to the same topic.
The University of Michigan, as an equal opportunity/affirmative action employer, complies with all applicable federal
and state laws regarding nondiscrimination and affirmative
action, including Title IX of the Education Amendments of
1972 and Section 504 of the Rehabilitation Act of 1973.
The University of Michigan is committed to a policy of equal
opportunity for all persons and does not discriminate on the
basis of race, color, national origin, age, marital status, sex,
sexual orientation, gender identity, gender expression, disability, religion, height, weight, or veteran status in employment,
educational programs and activities, and admissions. Inquiries
or complaints may be addressed to the Senior Director for
Institutional Equity and Title IX/Section 504 Coordinator,
Office of Institutional Equity, 2072 Administrative Services
Building, Ann Arbor, Michigan 48109-1432, 734-763-0235,
TTY 734-647-1388. For other University of Michigan information call 734-764-1817.
The Regents of the University
Julia Donovan Darlow, Ann Arbor
Laurence B. Deitch, Bingham Farms
Denise Ilitch, Ann Arbor
Olivia P. Maynard, Goodrich
Andrea Fischer Newman, Ann Arbor
Andrew C. Richner, Grosse Pointe Park
S. Martin Taylor, Grosse Pointe Farms
Katherine E. White, Ann Arbor
Mary Sue Coleman, ex officio
Crash Protection for Child Passengers:
Rationale for Best Practice
by Kathleen D. Klinich, Miriam A. Manary, and Kathleen B. Weber
Vehicle Crashes. . . . . . . . . . . . . . . . . . . .
Restraint Systems. . . . . . . . . . . . . . . . . .
Seating Position, Airbags, and Children. . .
Seatbelt. . . . . . . . . . . . . . . . . . . . . . . . .
LATCH . . . . . . . . . . . . . . . . . . . . . . . . . .
Usability and Vehicle/Child
Restraint Compatibility . . . . . . . . . . . . . .
CHILD RESTRAINT SYSTEMS . . . . . . . . . . . . . . 9
Rear-Facing Child Restraints. . . . . . . . . . . 9
Car Bed Restraints. . . . . . . . . . . . . . . . . 14
Forward-Facing Child Restraints. . . . . . . 15
CHILD RESTRAINT TESTING. . . . . . . . . . . . . .
Test Procedures. . . . . . . . . . . . . . . . . . .
Injury Criteria Limitations . . . . . . . . . . .
Dummy Limitations. . . . . . . . . . . . . . . .
CHILD BOOSTERS AND BELTS. . . . . . . . . . . .
Boosters. . . . . . . . . . . . . . . . . . . . . . . .
Boosters and Lap Belts. . . . . . . . . . . . . .
Seatbelts for Children . . . . . . . . . . . . . .
hild restraint systems provide
specialized protection for pediatric motor vehicle occupants whose
body structures are still immature and
growing. There are many occupant
protection systems available and the
different types of restraints are appropriately matched to children based
primarily on their ages and sizes.
Even with the most appropriate child
restraint, the way in which it is installed
and used can affect its performance.
This review describes the basic principles behind the design of occupant
restraint systems and applies them to
the needs of children. It also includes a
brief overview of child restraint testing
procedures and their limitations. Each
section describes research and insight
behind current best practice concepts,
primarily from the US perspective.
CONCLUSION. . . . . . . . . . . . . . . . . . . . . . . 28
ACKNOWLEDGMENTS. . . . . . . . . . . . . . . . . . 28
REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . 29
Child restraint systems are highly
effective safety devices that protect children against the leading cause of injury
and death for those ages 3–18 years.
There are four main sequential steps in
child restraint best practice: rear-facing
harnessed restraint, forward-facing
harnessed restraint, booster seat with
a lap-and-shoulder belt restraint, and
seat belts. Matching the appropriate
restraint mode with the child depends
on their physical size and maturity level.
Relative to no restraint, rear-facing
child restraints are 71% effective and
forward-facing child restraints are 54%
effective in reducing the likelihood
of death and serious injury (Kahane
1986). Forward-facing child restraints
reduce odds of an injury by 78–82%
compared with lap and shoulder belts
(Zaloshnja et al. 2007, Arbogast et
al. 2004). Children up to age 2 in
forward-facing child restraints are 1.76
times more likely to experience serious
injury than children in rear-facing child
restraints, with the analysis showing
a more distinct benefit for rear-facing
children in side impact crashes (Henary
et al. 2007). Children aged 2–6 using
child restraints without gross misuse
had a 28% lower risk of death than children using only seatbelts (Elliott et al.
Use of a booster seat reduces risk
of fatal injury for children aged 4–8
years by 55–67% relative to the risk to
unrestrained children and adults, while
UMTRI Research Review
Researchers at the Children’s
Hospital of Philadelphia in partnership with State Farm Insurance
Companies conducted a study
called the Partners for Child
Passenger Safety (PCPS) from
1998 through 2007. Crashes
involving children in 15 states
and the District of Columbia were
identified through State Farm
Insurance claims, and data on
the crashes were collected using
a validated telephone survey and
selected in-depth crash investigations. The survey identified
clinically significant injuries, which
generally correspond to those
AIS2+ and higher such as broken
bones and organ injuries. The
complete dataset includes 34,732
crash-involved children under age
16 who were in 21,943 crashes.
Using statistical techniques to
develop weighting factors to
represent crashes in the US population, these children’s outcomes
can be analyzed to represent
531,193 children in 346,485
crashes. This dataset provides
valuable information for evaluating actual performance of child
restraint models available up to
2007 in real crashes, and a significant portion of the recent data
included in this review comes from
this study.
January - March 2012
children aged 5–14 using seatbelts have
a 52% reduction in fatal injury risk
relative to unrestrained occupants’ risk
(Morgan 1999). While proper booster
seat use has only a marginal effect on
reducing risk of fatal injury compared
with the risk of using a seatbelt alone,
children aged 4–8 using booster seats
have a 55% lower risk of serious injury
compared with those using seatbelts
alone (Arbogast et al. 2009). In
comparison, seatbelts are 37–48% effective at preventing fatal injuries in adults
while seatbelts plus frontal impact
airbags are 44–54% effective at preventing fatal injuries (Morgan 1999). While
child restraints are the most protective
restraints available, there can still be
catastrophic crashes, particularly those
involving severe intrusion into the
passenger compartment of the vehicle,
where child restraints are not effective
and children can sustain severe or fatal
Tremendous strides have been
made to increase child restraint use
nationwide, primarily through the
enactment of state child passenger
safety laws and educational campaigns
aimed at parents and caregivers. Since
2000, 49 of 50 states have improved
their child restraint laws either by
extending the applicable child ages,
making nonuse a primary enforcement
infraction, or adding other best practice
elements to the law. Over the past 25
years, the proportion of children riding
unrestrained has decreased from 90%
in 1976 (Williams 1976) to approximately 13% in 2008 (NHTSA 2009).
However, unrestrained children ages
0–15 years account for approximately
54% of children killed each year in
motor vehicle crashes (NHTSA, 2002),
indicating that additional gains in safety
can be achieved by minimizing nonuse
of child restraints further.
Misuse of child restraints remains
common, with estimates of misuse
ranging from 63–90%. (Decina and
Lococo 2005 and 2007, Eby and
Kostyniuk 1999, Dukehart et al. 2007)
Some errors may have minimal effect
on safety, particularly in less severe
crashes. However, multiple “small”
errors may combine to cause as much
of a decrement in performance as a
single major error, and become more
critical as the severity of the crash
increases (Tsai and Perel 2009, Lesire
et al. 2007, Weber and Melvin 1983).
The effectiveness of child restraints is
estimated from their “as used” state,
which includes misuse conditions, and
thus would be expected to increase if
misuse were reduced.
Vehicle Crashes
Vehicle crashes consist of a series
of collisions; the most common type
is a frontal impact. The initial impact
is between the vehicle and another
object, while the occupants continue to
move forward at the precrash speed as
the vehicle slows down. Unrestrained
occupants will continue at the precrash
speed until they abruptly stop against
the decelerating vehicle interior or
surfaces outside the vehicle, experiencing high accelerations and loading.
Restrained occupants collide with their
restraints, giving them a longer time
and distance to come to a stop so that
they experience lower loads and acceleration levels.
The front structures of vehicles
are designed to crush during frontal
crashes, thereby absorbing a portion
of the crash energy and allowing the
passenger compartment to stop over a
greater distance and longer time than
does the front bumper. By coupling the
occupants to the passenger compartment structure with snug-fitting belts,
they will “ride down” the crash with
the vehicle frame for a longer time
period. For small children to “ride
down” the crash, a snug harness
couples them to the child restraint, and
a tight installation of the child restraint
(using seatbelt or LATCH) couples
them to the vehicle.
In other crash directions, the occupant motion is primarily toward the
point of vehicle impact. Although side
impacts usually have a lower change in
velocity than frontal impacts, there is
much less vehicle structure available to
absorb energy between the occupant
and striking object. Rear impacts are
generally the least severe among all
crashes, with restraint provided by the
vehicle seat back and head restraint.
Rollovers involving more than onequarter turn of the vehicle project the
occupant towards the roof, making
restraint use critical to achieiving good
injury outcomes.
Restraint Systems
Vehicle seat belts or harnessed
restraints that are initially snug allow
immediate restraint of the occupant,
which maximizes the time of restraint
and minimizes the level of loading
required to stop the occupant. Other
supplemental protection systems, such
as padding or airbags, can absorb
impact energy between the occupant
and the vehicle interior. If belt or
harness webbing is loose, the occupant
will travel farther before restraint can
begin, increasing the level of force
needed to stop the occupant in a
shorter time period. Advanced seat belt
designs balance between loading the
occupant and controlling contact with
vehicle interior components.
To optimally reduce the risk of
injury, the remaining loads must be
distributed as widely as possible over
the body’s strongest components. For
adults who face forward, these parts
include the shoulders and pelvis. For
children, especially infants, distributing the restraint loads over larger and
sometimes different body areas is
necessary. Multiple straps and rearwardfacing orientation help take care of
these needs.
The primary goal of any occupant
protection system is to keep the central
nervous system from being injured.
Broken bones will mend and soft tissue
will heal, but damage to the brain and
spinal cord is typically irreversible. In
the design of restraint systems, tradeoffs
may be necessary that compromise on
protection for the extremities or ribs
to ensure protection of the brain and
spinal cord.
Proper belt placement and good
fit are important for effective seat belt
restraint when using either the vehicle
seat belt or a child restraint harness.
Serious restraint-induced injuries can
occur when the belts are misplaced over
body areas having no protective bony
structure. Such misplacement of a lap
belt can also occur during a crash if the
belt is loose or, with small children,
is not held in place low on the pelvis
by a crotch strap or other positioning
device, such as booster belt guides. A
lap belt that is placed or rides up above
the pelvis can intrude into the soft
abdomen and rupture or lacerate internal organs (Rouhana 1993, Rutledge et
al. 1991). Moreover, in the absence of
In the mid 1990s, NHTSA
began support of a national standardized child passenger safety
curriculum that has created a
nationwide network of specially
trained personnel to educate
parents and caregivers about
correct practices in child passenger
safety. The course, implemented
in 1998, consists of 3-5 days of
instruction and demonstrations,
culminating in a community car
seat check up event. Students
demonstrate their mastery of the
material with written test and skills
demonstrations and then must
recertify every two years. Since
its introduction, the program has
trained over 108,000 technicians,
and about 30,000 are currently
certified. This group includes
first responders, law enforcement
personnel, firefighters, parents,
researchers, clinicians and social
workers. These trained technicians
can use their skills to staff permanent fitting stations and car seat
check up events as well as serving
as community educators.
For more information see
UMTRI Research Review
Childhood obesity leads to
many health problems, but the
data are mixed as to whether there
is a potential for increased injury
rates in motor-vehicle crashes
(Haricharan et al. 2009, Zonfrillo
et al. 2011). A primary concern is
that children may not be able to
follow best-practice recommendations for their age (rear-facing
until 2 years, harnessed restraint
to at least age 4, booster seat until
age 8-12 when the seatbelt fits
correctly) because they exceed
the weight limit of the available
products (Trifiletti et al. 2006,
Fitzharris et al. 2008, Bahlmann
et al. 2009). For example, a
three-year-old weighing 36 kg
(80 lb) would be too heavy for
nearly all forward-facing harnessed
restraints, but probably would
not sit correctly in a booster
that allows increased freedom of
movement. For children using
child restraints with an acceptable
weight limit, the dimensions of
the product or the harness system
may not fit them appropriately.
Another challenge is that pediatric crash dummies represent
average sizes of children, and
cannot identify how an obese child
may interact differently with the
January - March 2012
a shoulder restraint, a lap belt routed
above the pelvis will compress the soft
tissue and organs of the abdomen and
load the spine, possibly causing separation or fracture of the lumbar vertebrae
in a severe crash (Johnson and Falci
1990, King 1993). Misusing a lap-andshoulder belt by placing the shoulder
belt behind the back removes torso
restraint and allows the same problems
seen with lap belts only; placing the
shoulder belt under the arm provides
minimal torso restraint and can increase
direct loading to the abdomen and
chest compared with a properly positioned belt (McGrath et al. 2010,
Louman-Gardiner et al. 2008).
Despite the potential for beltinduced injuries, belt-based restraint
systems have significant advantages
over supplementary airbag systems.
They offer protection in a variety of
crash directions, including rollovers,
and throughout the course of multiple
impacts. Moreover, the force on the
occupant is proportional to the mass
of that occupant. For example, a man
weighing 80 kg will experience a much
greater restraint load into the belts on
his chest and pelvis than a child weighing only 20 kg. Even though the child’s
bony structure and connective tissue
may be weaker than the adult’s, the
child’s mass is so much less that the
injury potential from contact with belts
or other restraint surfaces is also less. In
contrast, first-generation frontal airbags
produce the same amount of deployment force and resistance to deflation
regardless of occupant size, while some
advanced airbag systems vary deployment force based on the weight of the
Child restraint designs vary with
the size of the child, the direction
the child faces, the type of internal
restraining system, and the method
of installation. All child restraints,
however, work on the principle of
coupling the child as tightly as possible
to the vehicle because it maximizes
the time restraint can be applied and
minimizes the highest level of force
required to stop the occupant. In
North America, the child restraint
has been traditionally attached to the
vehicle with the existing seatbelts. An
option available in the US since 2002
is the LATCH system, which stands
for Lower Anchors and Tethers for
CHildren. After installing the child
restraint to the vehicle, the child is
then secured in the child restraint with
a separate harness. This results in two
links between the vehicle and the occupant. It is therefore critical that the
seatbelt or LATCH strap be tight and
the harness be snug to allow the child
to ride down the crash with the vehicle.
Seating Position, Airbags, and
From the early days of child
restraint regulation, the center rear seat
position has been considered the safest
place in the car, since it is farthest from
the exterior of the vehicle. Analyses of
field injury data continue to bear this
out (Kallan et al. 2008, Braver et al.
1997, Mayrose and Priya 2008). Before
the implementation of frontal-impact
airbags for the right-front passenger,
infants were often restrained in the
front seat to allow monitoring by the
driver (Edwards and Sullivan 1997).
In addition, until lap-and-shoulder
belts were required in rear outboard
seats in 1989, the front seat offered
the only passenger position with the
more complete lap-and-shoulder belt
When frontal-impact airbags
were required in the United States to
provide protection for unbelted rightfront passengers in the early 1990s, the
unintended consequence of a restraint
system designed for adults was the
potential for lethal loading of children
riding in the front seat. (QuinonesHinojosa 2005, Braver et al. 1998). To
date, 28 infants and 152 older children
sustained fatal loading attributed to the
airbag in the United States (NHTSA
2009). The increased risk of fatality to
children in the right-front passenger
seating position in vehicles with firstgeneration airbags is estimated to be
34%–63%. (Braver et al. 1998). These
fatalities almost always involved head or
neck injury from direct contact with the
inflating bag and/or the airbag housing
cover delivered to children who were
either riding in a rear-facing child
restraint or unrestrained and/or were
out of position and close to the airbag
at the instant of deployment (NHTSA
The immediate response to the
injuries and fatalities to children by
first-generation airbags was to recommend that children under 13 years of
age use the rear seat. The combination
of airbag warnings on child restraints
and in vehicles along with educational
campaigns has led to high use of the
rear seat by children, increasing to 83%
of children under age 7, including most
children in rear-facing restraints and
50% of children aged 8–12 (Greenspan
et al. 2010). Another study indicates
that 99% of children under age 1,
98% of children aged 1–3, and 88% of
children aged 4–7 rode in rear seats
in 2008 (NHTSA 2008). In addition,
vehicles are now equipped with occupant detection systems that are meant
to automatically turn off the airbag and
prevent frontal-airbag deployment with
child occupants and occupants who
are too close to the deploying airbag.
Federal testing requirements have also
changed so airbags deploy with less
force. Braver et al. (2008) showed
that relative to first-generation airbags,
second-generation airbags led to reductions in fatal injuries of 65% for children
aged 0–4, 46% for children aged 5–9,
next to the airbag module (Side Impact
Working Group 2003). Curtain airbags
deploy from the vehicle roofline and
provide head protection during side
impacts and rollovers. Field investigations of crashes have identified almost
no unintended injuries to children
caused by side or curtain airbags, indicating that the efforts to ensure the
safety of their implementation have
been effective (Hallman et al. 2009,
Arbogast and Kallan 2007).
and 32% for children aged 10–12 who
were seated in the front seat. Olson
et al. (2006) found a 34% reduction
in risk for children under 6 between
second- and first-generation airbags.
Regardless of what type of airbag
system exists in the vehicle, children
under age 13 should ride in the back
seat (Arbogast et al. 2009, AAP 2011,
NHTSA 2011). Despite the improvements in airbag technology, this
recommendation remains important
because the vehicle fleet still includes
vehicles where the airbag can pose a
danger. In addition, the rear occupant
compartment provides a safer environment during a frontal crash because of
intrusion that is more likely to occur
in the front occupant compartment
(Evans et al. 2009). Several studies have
documented the protective effect of the
rear seat for belted occupants (Berg et
al. 2000, Durbin et al. 2005).
As side impact airbags have been
introduced to the vehicle fleet, more
precautions have been taken to avoid
the unintended dangerous consequenc–
es experienced with frontal-impact
airbags. Voluntary testing procedures
used by vehicle manufacturers evaluate whether the side airbags pose a
danger to an “out-of-position” child
Seatbelt use increased nationwide
from 11% in 1982 to 85% in 2010
(NHTSA 2009, Lund 1986), largely
due to the enactment and enforcement
of state occupant restraint usage laws.
During this time, vehicle manufacturers
developed seatbelt designs to improve
comfort, ease-of-use, and protection for
adult occupants. Some of the improvements for adult seat belts conflicted
with easy securement of child restraints.
For example, a seatbelt anchorage
located forward of the vehicle seat
bight (the intersection of the seatback
and bottom cushion) can provide
a more advantageous angle for belt
restraint of an adult, but makes it challenging to tighten a seatbelt adequately
during child restraint installation. As
a result, some issues regarding child
restraint compatibility with vehicle
belts and seats were addressed in SAE
Recommended Practice J1819 (1994)
and by the addition of a seatbelt lockability requirement to FMVSS 208 in
A common child restraint misuse
with seatbelt installations is the failure
of the installer to lock the seatbelt.
Many seatbelts are equipped with an
emergency locking retractor, which
locks to prevent forward movement
of the occupant during a crash event
UMTRI Research Review
Figure 1. Vehicle with visible lower anchorages.
January - March 2012
In response to the challenges
posed with seatbelt installation of
child restraints, the National Highway
Traffic Safety Administration (NHTSA)
introduced a new child restraint securement system in 1999. The Child
Restraint Anchorage System, commonly
called Lower Anchors and Tethers for
Children, is known as LATCH in the
United States and is defined in FMVSS
225 and additions to FMVSS 213. The
LATCH concept originated from an
effort in the International Standards
Organization (ISO), which proposed
and adopted a universal child restraint
anchorage system called ISOFIX (ISO,
1999a). Implementation of LATCH
in the United States began in 1999
and was required in vehicles and child
restraints in 2002.
The ISOFIX concept calls for
two lower attachment points and a
means to “limit pitch rotation of the
child restraint”. In the United States,
LATCH has two distinct components:
lower connectors on child restraints
that attach to lower anchorage points
located at the vehicle seat bight (figure
1), and a top tether strap on forwardfacing restraints that attaches to anchor
points located on the rear shelf, seat
back, floor, cargo area, or ceiling of the
vehicle (figure 2).
Most US child restraints are
equipped with a LATCH strap consisting of a length of webbing with
adjustment hardware and connectors
Figure 2. Sedan in which the tether anchorage is located under a
marked door on a package shelf (shown closed and open).
but allows movement of the occupant
during normal driving. Use of a seatbelt
equipped with an emergency locking
retractor alone allows the child restraint
to shift during normal driving. To
eliminate this problem, some retractors
are switchable, and can be converted to
an automatic locking retractor, which
allows the belt to be locked tightly
through a child restraint belt path.
They are usually switched by pulling the
webbing completely out of the retractor
and then feeding it back in to tighten.
Other belt systems use a locking
latchplate that allows tight installation
of the child restraint in most cases,
although it can sometimes be incompatible and loosen during use. Some
child restraints are equipped with belt
lockoffs that can lock the seatbelt
by clamping down on the webbing
without use of vehicle hardware.
Locking clips are still provided on child
restraints without belt lockoff hardware
and can be used to prevent transfer of
webbing from the shoulder portion to
the lap portion of the belt. However,
locking clips are often ignored or
placed incorrectly (instead of properly positioned within one inch of the
latchplate), which may cause them to
deform, fly off, and/or introduce belt
slack during a crash.
A new challenge in obtaining tight
installation of rear-facing restraints
using seatbelts has become more
prevalent with the presence of lapand-shoulder belts in all rear seating
positions. A tight lap-and-shoulder belt
can cause a rear-facing child restraint
to tilt sideways as the taut shoulder
belt portion pulls up on the belt path.
If needed, a child restraint lockoff or a
locking clip can be used to allow tightening without tipping.
on each end. The two most common
types of connectors are hook-on and
push-on (figure 3). The LATCH strap
is usually routed through the appropriate belt path on the child restraint that
would also usually be used to route the
seatbelt (figure 4 top) or attached to
each side of the child restraint (figure 4
center). The LATCH strap is designed
to replace the vehicle seat belt as the
primary securement system. Rigid
lower LATCH connectors have been
manufactured on a few models of US
child restraints (figure 4 bottom), but
are widely used in Europe where they
are required for ISOFIX. Attaching
the top tether achieves a more secure
installation and reduces occupant excursions when installing a forward-facing
restraint with either the LATCH strap
or vehicle seat belt. While using the
tether improves occupant protection,
child restraints in the United States
must also pass less-stringent head excursion requirements without the tether
to ensure reasonable protection if the
caregiver fails to use it.
While many vehicles do allow easier
child restraint installation with LATCH
compared to seatbelts, in other vehicles
the interface with the LATCH hardware makes child restraint installation
difficult, and outright incompatibilities
between child restraints and particular vehicles have been documented
(IIHS 2003, SafeRideNews 2010).
New types of misuse have been identified when using LATCH. Top tethers
are only used about half the time,
even though all vehicles and restraints
have had ready-to-use tether hardware since 2001 (Decina and Lococo
2007, Jermakian and Wells 2010).
Errors in attaching tethers include
connecting them to the wrong hardware, misrouting them with respect to
the head restraint, connecting them
upside-down, and not tightening them
sufficiently. Errors in attaching lower
connectors include connecting them to
the wrong hardware, connecting them
upside-down, and failing to tighten
the webbing after connecting. In addition, installers often install the child
restraint using both the seat belt and
lower LATCH strap, which is only
currently allowed by one vehicle manufacturer. Because most US products
use the same belt paths to route either
manufacturer currently includes any
information about weight limits in their
vehicle manuals (Ford 2011). While
weight limits on lower anchorages may
be appropriate because there is another
means (seatbelt) to serve as the main
method of attaching the child restraint,
setting weight limits on tether use
seems misguided because it is a supplement to the main attachment system,
and the demonstrated benefits of top
tether use in reducing head excursion in
a wide range of crashes are much higher
than the possible risk of injury caused
by hypothetical tether anchorage failure
in an extremely severe crash.
Figure 3. Hook-on (top) and push-on (bottom) LATCH strap
the LATCH strap or the seatbelt, the
LATCH strap can also be misrouted
through the belt paths on the child
In some vehicles, LATCH has
fulfilled the intended goal of making
child restraint installation easier, thus
reducing misuse and improving effectiveness of the child restraint. However,
because of problems in some vehicles,
it may still be easier to achieve a better
installation using the seatbelt. Best
practice dictates that the easiest method
providing a tight installation should be
used to install a child restraint, keeping
in mind that the tether should always
be used for all forward-facing installations.
FMVSS 225 specifications include
lower and tether anchorage strength
requirements evaluated with a quasistatic pull test. When LATCH was first
implemented, most harnessed child
restraints could only accommodate
children up to 18 kg (40 lb). Since
then, a number of products have been
introduced that allow children up to
23, 29, or even 39 kg (50 lb, 65 lb, 85
lb) to use a harnessed restraint system.
Since there is no straightforward way to
identify the dynamic strength limits of
vehicle anchorages from the quasi-static
test data, some vehicle manufacturers have expressed concern that their
LATCH hardware should not be used
with harnessed child seats for larger
children. NHTSA clarified LATCH
strength issues in a regulation stating
that they consider the strength of lower
and tether anchorages (based on static
testing) sufficient to secure a child and
child restraint with a combined weight
of 65 lb (NHTSA 2012). Although
many vehicle manufacturers have
provided recommended weight limits
(that are lower than this value) to a
supplementary manual on LATCH
used by child passenger safety technicians (SafeRideNews 2011), only one
Figure 4. Implementations of LATCH on US child restraints: LATCH
strap routed through belt path (top), attached to bar on each side
(center), rigid LATCH (bottom).
UMTRI Research Review
January - March 2012
specific combinations of the two (ISO
2010, Pedder and Hillebrandt 2007).
This rating system currently focuses
on ISOFIX (LATCH-type) systems.
Some of the vehicle features that are
rated in the current version of the ISO
document include the vehicle owner’s
manual instructions on how to identify
the number and location of seating
positions available for child restraint
installation, the visibility and labeling
of the LATCH anchors, the presence
of other hardware elements that could
be mistaken for LATCH anchors, the
actions required for preparing the
seating position for child restraint
installation, and conflicts between
LATCH and seatbelts.
School buses are the safest
form of motor vehicle transportation, and recent data based on
fatality rates show that they are
eight times safer than riding in
typical passenger vehicles (NHTSA
2002). The high level of safety is
due in part to the special vehicle
construction requirements, the
high vehicle mass, the conspicuous
yellow color, their use primarily
during daylight hours on known
routes, and the extra training
bus drivers receive. School buses
were originally designed to transport children ages 6 and older to
school. They provide occupant
protection using a concept called
compartmentalization where the
closely spaced seats with extra
padding on the seatbacks create a
padded compartment to protect
the riders. Because federal funding
for Head Start requires many
children younger than 6 to be
transported in school buses using
child restraints, buses have some
seating positions equipped with
either seat belts or LATCH lower
anchors. School bus seating manufacturers and NHTSA state that
these seats with lap-and-shoulder
belts are built to fit a 6-year-old
child properly without a booster.
Therefore boosters are not recommended at all in school buses.
Other publications address details
about using child restraints on
school buses (SafeRideNews 2009).
Usability and Vehicle/Child
Restraint Compatibility
As implementations of LATCH
hardware in vehicles and on child
restraints have evolved over the past
decade, problems with incompatibilities
between vehicles and child restraints
remain. Caregivers also make mistakes
when securing their children in the
child restraint harness. Several different
rating systems have been proposed to
improve the usability of child restraints,
reduce misuse, and increase compatibility between the child restraint and
vehicle. Except for some label and
instruction issues, usability is not an
explicit part of FMVSS 213.
NHTSA developed an Ease-of-Use
(EOU) Rating system (NHTSA 2006)
to provide consumers with information about which child restraints have
features that enhance usability. The
system has provided strong incentives for child restraint manufacturers
to improve products, labeling, and
instruction manuals with respect to
usability. The rating system includes
questions that address each child
restraint area related to the most
common misuse modes. NHTSA has
also proposed a voluntary vehicle/
child restraint fit evaluation program
to encourage vehicle manufacturers
to provide information to consumers about compatibility for vehicle/
child restraint pairings (NHTSA 2010).
Vehicle manufacturers would submit
lists of child restraints that are compatible with a particular vehicle based on a
number of key installation factors.
In the field, some misuse modes
arise from features and elements of the
vehicle environment and others result
from interactions between specific child
restraint and vehicle combinations. A
usability rating scheme has been issued
by the ISO Child Restraints Group that
has rating forms for all three elements:
the child restraint, the vehicle, and
The SAE Children’s Restraint
Systems Standards Committee has
drafted a new recommended practice
to improve compatibility between child
restraints and vehicles during LATCH
installations (SAE 2007). The document defines tools and procedures for
evaluating hardware in vehicles and
on child restraints to improve their
ease-of-use. Factors include measuring
the force required to attach LATCH
strap connectors to lower anchorages,
measuring the clearance around lower
anchorages, and recommendations for
maximum size of LATCH connector hardware. Since the recommended
practice is still in draft form, vehicle
and child restraint manufacturers are
likely not using the proposed methods
There are four main sequential
steps in child restraint best practice:
rear-facing harnessed restraint, forwardfacing harnessed restraint, booster seat
with a lap-and-shoulder belt restraint,
and seat belts. Matching the appropriate
restraint mode with the child depends
on their physical size and maturity level.
Each successive step offers less occupant
protection than the previous mode, so
caregivers should delay moving to the
next level as long as possible.
Rear-Facing Child Restraints
Both the American Academy of
Pediatrics and the National Highway
Traffic Safety Administration now
recommend that children remain
rear facing until they outgrow their
restraint. This means that most children
can remain rear-facing through age
2 years, based on average child sizes
and the capacity of most rear-facing
convertible restraint products on the
market. US crash data show that children aged 1–2 years are 5.53 times
safer in a rear-facing restraint than in a
forward-facing restraint in side impacts
and 1.23 times safer in frontal impacts
(Henary et al. 2007). These recent US
data support Swedish data showing
benefit for children rear-facing through
age 4, with rear-facing restraints reducing AIS2+ injury by 90% compared
with unrestrained children (Jacobsson
et al. 2007, Isaksson-Hellman et al.
1997). Because earlier rear-facing child
restraints did not accommodate larger
children, older education materials may
contain outdated information stating
that children can begin using forwardfacing restraints at age 1 or 10 kg (20
lbs), which is no longer considered a
safe practice.
Types of rear-facing restraints
Two types of restraints, infant
restraints and convertibles installed
rear-facing, are commonly used to
orient the child to face the rear of the
vehicle. Infant restraints can only be
used rear-facing and most have a separate base which remains in the vehicle
to facilitate repeated installation (figure
5 top), but most can also be used
without the base and secured with the
seatbelt (figure 5 middle). The infant
restraint base can be installed with
either lower anchorages or the seatbelt.
These products usually have a carrying
handle. Traditionally, these products
have accommodated children up to 9 or
10 kg (20 or 22 lb), but there are now
many models that can accommodate
children up to 13 kg (30 lb).
A rear-facing convertible is shown
in figure 5 (bottom). These can be used
rear-facing up to 13–20 kg (30–45
lb), then converted for forward-facing
use. Convertibles tend to be larger
than infant seats. While most children will outgrow their rear-facing
restraint because they reach the allowable maximum weight limit for their
use, some children will outgrow their
rear-facing product because the top of
their head is within 2.5 cm (1 inch)
of the top edge of the child restraint
back support. It is common practice to
use an infant restraint for a newborn
until it is outgrown by weight or seated
height. For a product with a 9 or 10 kg
(20 or 22 lb) weight limit, this means
that most children would outgrow the
device within the first year and should
then be moved into a convertible child
Figure 5. Infant seat with base installed with LATCH (top);
infant seat without base installed with belt (center);
rear-facing convertible installed with LATCH (bottom).
restraint used rear-facing.
Rear-facing restraints use an internal harness to secure the child into the
shell. In a frontal impact, the restraint
forces occur where the back of the child
meets the restraint so that the restraining load is distributed across the entire
back and head of the infant.
UMTRI Research Review
Ambulances pose special challenges for correct child restraint
installation. Often there are no
forward-facing vehicle seats within
the transport cabin where a child
restraint can be installed per the
manufacturer’s instructions. If the
child needing transport is uninjured, the best practice approach
would be to call for a conventional
passenger vehicle with an appropriate child restraint to transport
the child. If the child is a patient
who requires care during transport, methods have been proposed
for securing certain types of child
restraints to the patient gurney
(Bull et al. 2001). Some ambulances are equipped with captain’s
seats that have built-in child
restraint systems.
The infant’s head is well supported
in this mode, and the movement of the
head and neck happen in unison with
the torso during a crash to eliminate
severe tension and flexion forces on the
neck that can occur with forward-facing
occupants. Figure 6 shows the differ-
ence in kinematics between the same
child restraint used rear-facing and
forward-facing in a simulated frontal
impact. Peak axial neck forces are four
times higher in the forward-restraint
compared to the rear-facing restraint.
Figure 6. During a frontal crash simulation (times 0, 60 ms, 90 ms), the back of a rear-facing restraint (farther from camera), supports
the head and back of the child. In the forward-facing child restraint (closer to camera), the head is pulled away from the restrained torso.
January - March 2012
Rear-facing restraint recline angle
The angle of installation is one
of the most critical factors for correct
restraint of children riding rear-facing.
If the restraint is too upright, newborn
infants may not be able to breathe
because their heads drop forward
during travel. If the restraint is too
reclined/flat, the child will not be
effectively restrained by the back of
the child restraint. Ensuring that the
child’s head is in contact with the child
restraint back support is also best for
crash protection.
Focusing on crash protection, if
the back support angle is more reclined
than 45°, the reaction force to restrain
the child in a frontal crash starts to be
exceeded by the force projecting the
baby upwards along the seatback and
toward the front of the vehicle. As the
child grows, gains weight, and can hold
its head erect, a more upright restraint
angle would provide better crash
protection (figure 8).
Figure 8. Back angles for optimal protection. A more upright,
30° angle is more protective for older children in a crash while
younger infants and those without good head control need a more
reclined support angle of 45°.
Figure 7. European child restraint with support leg.
Using a rear-facing infant restraint
facing forward can result in dangerous
loading and possible ejection because
the belt path has not been designed for
loading in this mode. Similar consequences could occur if a convertible
restraint is installed rear-facing using
the belt path for forward-facing (or vice
versa). Restraining an infant or toddler
forward-facing too early increases the
risk of injury to the spinal cord as the
child’s disproportionately large/heavy
head is stopped from forward motion
by a tension load applied in the cervical
Regulatory tests differ globally
with regard to the extended rear-facing
position. In Europe, dynamic tests for
both frontal and rear impacts (R44/04)
require additional attachment and stabilization elements to conform to the
requirements. Swedish rear-facing child
restraint designs differ from US products, in that they often use a support
leg (figure 7) and strap attachment
to the front seat or are placed in the
front seat against the instrument panel
with the airbag deactivated to limit
forward rotation of taller or heavier
children. These extra requirements are
less known in EU countries outside
Sweden, which may lead to a higher
risk of incorrect installation. In addition, these seats are approved for use in
specific vehicles, not across all models.
For the youngest infants, providing the best crash protection must be
balanced with providing an angle that
prevents the head from flopping over
and potentially pinching off the airway.
A back support angle of 45° from vertical is considered the maximum angle
that can achieve these two aims. To
Children with special medical
needs also require effective occupant protection. The same general
occupant protection principles apply,
such as rear-facing as long as possible, tight installation of the child
restraint, and snug harness adjustment. Given the needs of these
children, however, sometimes their
occupant protection system must
be different or include additional
postural support elements. Car
beds are one example, but there are
other systems to address most of the
commonly encountered healthcare
issues, including children in hip and
body casts, those with tracheotomies or muscle tone abnormalities,
and those who use wheelchairs for
ambulation. A national curriculum
titled “Safe Travel for All Children:
Transporting Children with Special
Health Care Needs,”
www. and the
American Academy of Pediatrics
(1999) offer additional information.
For children who use wheelchairs and cannot transfer to a child
restraint, the best practice is to use
a wheelchair that meets a voluntary
crashworthiness standard (RESNA
WC19), which means it is designed
to perform as a motor vehicle
seat. The wheelchair is attached
to the vehicle with a crash tested
securement system, most often a
four-point tiedown that complies
with RESNA WC18. The child in
the wheelchair must be provided
with a crashworthy belt system that
is properly fitted to their body. Most
belts attached to the wheelchair are
meant for postural positioning and
may not protect a child during a
crash. For more information about
wheelchair transportation safety, see or
UMTRI Research Review
January - March 2012
Tethering rear-facing
Following practices common in
Australia and Scandinavia, some rearfacing child restraints in the United
States provide the option to use a
tether to help secure the child restraint
to the vehicle, although all rear-facing
child restraints sold in the United States
need to meet the federal regulation
without a tether. FMVSS 213 does not
include any testing of a rear-facing child
restraint with a tether, nor does FMVSS
225 cover tether anchors that may
need to be located forward and below
a vehicle seating position for use with a
rear-facing child restraint.
the methods of rear-facing tethering,
any type of rear-facing configuration
(no tether, Australian tether, or Swedish
tether) provided superior protection
compared to forward-facing restraint
with tether. Among rear-facing tethering options, the Australian tethering
method produced the lowest accelerations and excursion to the dummy head
and chest among the methods evaluated. None of the tethering methods
produced potentially injurious neck
loads, based on the neck loading levels
established in FMVSS 208, during the
rear impact test events (Manary et al.
account for the differences in vehicle
seat angle, child restraint manufacturers often provide a means to indicate
and adjust the installation angle. At
least one major child restraint manufacturer sets its target angle at 35° from
vertical through the use of a visual
indicator, while others specify angles
closer to 45°. Unfortunately, indicators
provided on rear-facing child restraints
are primarily based on the angle where
the restraint performs best in regulatory
testing. Manufacturers may not fully
consider that the angle providing the
best orientation for a newborn may not
be the best choice for a larger toddler.
If the installation angle required by
child restraint instructions places a
newborn too upright, either a different
rear-facing child restraint or a car bed
should be used.
If a rear-facing restraint is installed
in a rear seat with its back initially
against the seat ahead, this will help
limit rotation during a crash and
provide improved protection, partly
because the child restraint will not
suddenly strike the seatback as it would
if there were an initial gap (Tylko 2011,
Sherwood et al. 2005). However, some
child restraint manufacturers prohibit
contact with the front seat because
of concerns about adverse interaction
between the child restraint and front
seatback in a rear impact. In some
vehicles with advanced airbags, vehicle
manufacturers also prohibit contact
between a rear-facing child restraint and
the right-front passenger seat because
it could interfere with occupant sensing
Figure 9. Tethering rear-facing child restraints: Australian method (left) and Swedish method (right).
The Australian tethering method,
shown in figure 9 left, routes the
tether rearward, towards the back of
the vehicle, to the standard tether
anchorage used for forward-facing
installations. This tethering more effectively limits forward rotation of the
restraint in a crash, minimizes movement into the front seat, and allows the
child to better ride down the crash with
the vehicle. The traditional Swedish
method, shown in figure 9 right, routes
the tether down and forward to a point
on the floor in front of the vehicle
seat. This approach helps adjust the
initial restraint angle and limits rotation towards the rear of the vehicle on
rebound (Sherwood et al. 2005).
In a laboratory study comparing
Rear-facing restraints in side impacts
and other impact directions
As in frontal impacts, the most
important priority for reducing injury
in side impact is to minimize or eliminate the head strike. If the child’s head
contacts something, it should be a
surface designed to absorb energy
and limit injury. A typical rear-facing
restraint will rotate toward the struck
side of the vehicle more than a forwardfacing restraint simply because of
the increased distance between the
combined center of mass of the occupied restraint and the belt path. Despite
this greater motion toward the intrusion, rear-facing child restraints are over
five times better at preventing injury in
side impacts than forward-facing child
Harnesses and fit
Most rear-facing child restraints
are now equipped with a five-point
harness, although the original designs
for rear-facing infant restraints were
usually equipped with a three-point
harness that did not include pelvic
straps. Premature and newborn infants
may be so small that many rear-facing
restraints seem too big. Manufacturers
have added lower shoulder harness
positions and greater harness adjustability to improve the fit for tiny
infants. Some child restraints come
with padded inserts that position the
infant’s body for improved harness
fit and offer lateral support, but are
removed for use with older children.
Padding that pushes the infant’s head
toward its chest should not be used. If
the infant’s head or body needs lateral
support beyond that provided by the
child restraint, padding can be placed
between the infant and the side of the
restraint. Firm padding, such as a rolled
towel, can also be placed between the
infant and the crotch strap to keep the
infant from slouching (AAP 2011a).
Supplemental thick, soft padding,
which has not been provided by the
child restraint manufacturer, should
not be placed under the infant, behind
its back, or between the infant and
the shoulder straps. Such padding will
compress during an impact, leaving
the harness loose on the infant’s body
of the child. If shoulder straps are positioned above the shoulders of a child
in a rear-facing restraint, the child can
slide up the seatback during impact
so the head is beyond the top of the
restraint, increasing risk of injury from
head contact. Smaller babies’ heads may
not reach the top of the restraint, but
they could experience higher loading
through the shoulders when stopped
against the shoulder straps. Loose
harnesses increase the chance of ejection and lead to increased loads once
the child begins loading the harness.
Use of a chest clip helps keep the
harness positioned on the shoulders but
cannot compensate for a loose harness
during a crash.
restraints (Henary et al. 2007).
Most side impacts also have a
frontal deceleration component so the
occupant usually moves toward the
front and side of the vehicle simultaneously. When this happens, the head
of the child in a rear-facing restraint
will be directed further within the
protection offered by the side wings
of the restraint. This differs from a
forward-facing child restraint, where
the child’s head tends to move forward
and around the sidewings and be more
vulnerable to injury from the intruding
vehicle or door structure. Several laboratory studies have demonstrated that
a more rigid installation between the
child restraint and vehicle, such as that
provided by rigid LATCH attachments,
also works better to keep any child
restraint in position and prevent the
head from contacting vehicle interior
components (Klinich et al. 2005).
In rear-end and rollover crashes,
the shoulder straps act to contain the
child within the rear-facing restraint,
which may rotate up against the vehicle
seatback. This motion was originally
touted as a benefit by the early designers to protect the infant from flying
debris (Feles 1970). Since most rearfacing restraints are now larger and
taller, this gives them greater potential
to allow contact between the child’s
head and interior vehicle components
in a rear impact or rollover. However,
injuries from this mechanism have not
been documented in the field.
and allowing increased sliding upward/
forward toward the front of the vehicle
and increasing the risk of occupant ejection.
In a rear-facing restraint, shoulder
straps should be routed to restraint
slots that are at or below the shoulders
Rear-facing restraints and frontal
Frontal impact airbags and rearfacing child restraints do not mix.
Even with advanced airbag systems,
rear-facing restraints should never be
installed in the right-front passenger
seating positions. Installing any type of
rear-facing child restraint in a seating
position with a frontal impact airbag
carries a high risk of injury or death
during a crash. Frontal impact passenger airbags are stored in the instrument
panel and need a certain amount of
space in which to inflate before they
begin to act as energy-absorbing cushions for larger occupants. A rear-facing
restraint in the front seat places the
child’s head and body very close to the
airbag hardware. When current airbags
deploy in a crash, whether severe or
moderate, they emerge in a small folded
package at very high speed—as much as
300 km/h (186 mi/h). If an airbag hits
the back of a rear-facing child restraint
while it is still inflating, it will strike
with considerable force.
Accelerations measured at the
heads of infant dummies in this situation range from 100 to 200 g, (Weber
1993, Klinich et al. 2002) with 50 g
UMTRI Research Review
considered the threshold for injury for
children represented by a 6-month size
dummy (Klinich et al. 2002, Melvin
1995). The sequence shown in figure
10 shows the initial impact of the airbag
into a rear-facing child restraint, which
laboratory measurements have demonstrated is the cause of fatal head injury
in crashes. Although the airbag could
also propel the infant and rear-facing
child restraint into the vehicle seatback,
the head injury from the airbag would
already have occurred with the initial
airbag contact into the back of the
Car Bed Restraints
For infants with documented
breathing problems or who cannot
otherwise tolerate the semireclined
positions, a car bed is a suitable alternative to a rear-facing infant restraint.
The three models currently available
in the United States accommodate
infants ranging from birth weight to
15 kg (35 lb). In a car-bed restraint
(figure 11), the infant lies flat, preferably on its back. The car bed is placed
on the vehicle seat, with its long axis
perpendicular to the direction of travel
and the baby’s head toward the center
of the vehicle (not next to the door).
Depending on the car bed model, the
infant can be placed on its back, which
Figure 11. Car bed restraint installed with seatbelt. This model is
equipped with a strap (gray) to help keep the seatbelt in place.
Figure 10. Laboratory reconstruction of
airbag deployment into rear-facing infant
restraint that resulted in infant fatality.
T0, 95, 100, 110, 120 ms
January - March 2012
is preferred, on the stomach, or on the
side. In a frontal crash, the occupant
restraint forces are distributed along
the entire length of the infant’s body,
while a harness or other containment
device keeps the baby in place during
rebound or rollover. In a side impact,
however, the infant’s head and neck are
more vulnerable in a car bed than in
a rear-facing restraint, especially if the
impact is on the side nearest the head
and there is significant intrusion (Weber
1990). Field data from the United
States and other countries are sparse
but have not revealed any protection
deficiencies with this configuration.
The American Academy of
Pediatrics prefers the use of the semi-
Forward-Facing Child Restraints
Types of forward-facing child
There are two main types of
harnessed restraint systems that face the
child toward the front of the vehicle.
One is a convertible child restraint
used forward-facing (figure 12 top).
The other is referred to as a combination child restraint. Combination seats
Figure 12. Convertible child restraint installed forward-facing
with LATCH (top) and combination restraint used with harness and
installed with seatbelt (bottom).
(figure 12 bottom) are initially used
with a harness; the harness is then
removed to convert the restraint into
a belt-positioning booster. In addition,
a few products have been designed for
only forward-facing harnessed use.
Historically, forward-facing
restraints were made for use with a
harness for a child up to only 18 kg
(40 lb). However, many current models
now accommodate children up to 30 to
40 kg (65 to 90 lb) using the harness
system. These higher-weight harness
systems may include higher slots for
routing the harness straps at or above
the shoulders of a larger child, as well
as higher seatbacks that need to extend
to a height at or above the child’s ears
to protect against rearward bending
of the neck. Both forward-facing child
restraint types are installed with a seatbelt or LATCH lower attachments.
In addition, all current forward-facing
child restraints recommend use of the
tether with any installation to reduce
head excursion during a crash, and
some manufacturers require tether use
for the heavier children.
Harnesses and shields
The ability of a forward-facing restraint
to provide effective protection depends
on harness fit and snugness as well as
tight coupling to the vehicle. Current
child restraints are equipped with a
five-point harness, although a few child
restraint models still secure the child
with a tray shield and shoulder straps
(figure 13). The five-point harness
styles are generally preferred because
they permit a snug fit around the child.
However, the tray shield style may be
helpful for caregivers with a lack of
dexterity who may be unable to appropriately buckle the harness.
The five-point strap harness
arrangement is generally styled after
military and racing harnesses. Straps
go over each shoulder and the lower
portions form a lap belt across the
reclined, rear-facing position, but
recognizes the issues of positional apnea
(Degrazia et al. 2010, Nagase et al.
2002). It currently recommends that
all infants born at less than 37 weeks
gestation be monitored in a semiup–
right position prior to discharge from
the hospital to detect possible apnea,
bradycardia, or oxygen desaturation
(AAP 1999b).
Figure 13. Child restraints that secure child using harnesses (top)
or tray shield (bottom).
thighs as two latchplates connect to a
central buckle. The buckle, which is on
the end of the crotch strap, is routed
between the child’s legs, and serves
to hold the lap straps down on top of
the thighs, so it should be as short as
possible. Most current products have
a single pull harness adjustment strap
or knob that makes it easier to tighten
the harness so it is snug around the
child compared with earlier designs.
Loose harness straps will allow the
child greater movement toward vehicle
UMTRI Research Review
interior surfaces and generate higher
loads on the child when the system
finally pulls up tight to resist movement. Failure to buckle the harness or
route the harness properly could result
in ejection or serious injury to thoracic
and abdominal organs.
The shoulder straps of the forwardfacing harness should be routed to shell
slots located at or above the child’s
shoulders. For forward-facing restraints,
erroneously placing the harness shoulder straps in slots located below the
shoulders has the consequence of
introducing slack in the harness, as the
child’s torso can move forward before
the straps begin restraining the shoulder
and also creating increased compression loading in the spinal cord. Using
harness slots not specified for forwardfacing use may lead to child restraint
shell failure, as some lower slots on
convertible restraints are not reinforced
for loading in frontal mode.
Neck injury in forward-facing child
A transition to a forward-facing
child restraint should not be celebrated
but delayed as long as possible. A child
“graduating” to facing forward actually
experiences a decrease in protection from riding rear-facing, which is
the safest mode of restraint available
for children. While education about
the benefits of extended rear-facing
restraint use has become more widespread, there are still misconceptions
even within the medical community
about the appropriate timing for the
transition to forward-facing restraints.
A forward-facing child with shoulders held back by a harness during a
significant frontal impact can experience
severe loading of the cervical spine as
the mass of head extends forward and
is stopped by the neck. In a 48 km/h
(30 mph) crash with a 25-g passenger compartment deceleration, for
January - March 2012
instance, the head of a forward-facing
adult or child may experience as much
as 60 or 70 g, because the occupant’s
head stops later in the event and more
abruptly than the vehicle’s floor pan.
Even the strong neck muscles of military volunteers make little difference
in outcomes in such an environment.
Rather it is the skeletal strength of the
vertebrae, in combination with the
tightness of the connecting ligaments,
that determines whether the spine will
hold together and the spinal cord will
remain intact within the confines of the
vertebral column (Huelke et al. 1992,
Stalnaker 1993). Adult cervical spines
can withstand severe tensile forces
associated with decelerations up to 100
g (McElhaney and Myers 1993) and
failure is nearly always associated with
vertebral fracture.
On the other hand, the immature
vertebrae of young children consist of
both bony segments and cartilage, and
the ligaments are loose to accommodate growth (Kumaresan et al. 1998,
Myers and Winkelstein 1995). This
combination allows the soft vertebral
elements to deform and separate under
crash conditions, leaving the spinal
cord as the only fragile link between
the head and the torso. This flexibility
allows children to sustain spinal cord
injury without fracture to the vertebrae, which is extremely rare in adults.
Mathematical models of pediatric spines
(age 1, 3, and 6 years) subjected to
various types of loading indicate that,
compared to adult spines, the anatomical and material properties of immature
spinal elements make them much more
flexible than would be predicted by
relative size alone (Kumaresan et al.
2000). Crash experience has shown that
a young child’s skull can be separated
from its spine by the force of a crash
(Fuchs et al. 1989), the spinal cord can
be severed (Hoy and Cole 1993) or
the child may live but suffer paraplegia
or quadriplegia due to the stretched
and damaged cord (Langweider et al.
1990, Trosseille and Tarriere 1993,
Weber et al. 1993). The risk of spinal
cord injury in children increases with
crash severity and decreases with age
(Stalnaker 1993). Although serious
cervical spine injuries are rare among
properly restrained forward-facing children, because of the potentially severe
consequences, best practice dictates
the relatively simple countermeasure of
restraining smaller children rear-facing
as long as possible (figure 6).
Tethers and crash performance
Top tethers should always be used
with forward-facing child restraints to
anchor the top of the child restraint to
the vehicle and reduce forward rotation of the child restraint in a frontal
crash (Brown et al. 1995, Legault et
al. 1997). Figure 14 shows a crash
sequence comparing the performance
of the same model of child restraint
tethered (closer to camera) and untethered (farther from camera) in a 48
km/h crash test with a 3-year-old
sized dummy. The dummy in the child
restraint attached with a tether experiences about 150 mm (6 inches) less
forward movement of the head.
Reduced head excursion means that
in an actual crash, a child would be less
likely to experience head contact with
the interior. Among children injured
in forward-facing child restraints,
head and facial trauma predominate
(Nance et al. 2010, Arbogast et al.
2002). Head contact while the neck is
in tension can also generate vertebral
fractures and dislocations, as well as
spinal cord injury, by suddenly stopping the free motion of the head and
putting significant compressive and
shear loads on the neck (Stalnaker
1993, McElhaney and Myers 1993).
Reduction of head excursion and elimination of head contact are therefore as
important for avoiding neck injury as
they are for reducing head and facial
injury in children. Top tethers can also
partially compensate for suboptimal
installation tightness using the LATCH
strap or seatbelt by improving coupling
between the child restraint and vehicle.
However, the tether must be tight,
as the improvement offered by a top
tether is also degraded by slack. Failure
to use the tether is a common misuse
of forward-facing child restraints,
occurring in half of forward-facing
installations (Jermakian and Wells
Figure 14. Crash sequence showing reduced head excursion
of 15 cm (6 inches) with tether use (near side) versus
untethered (far side) child restraint.
Forward-facing child restraints and
side-impact protection
Although frontal impacts are the
most common type of crash, side
impacts are more likely to result in
serious and fatal injuries (Viano and
Parenteau, 2008). Rear-facing restraints
are so much more effective in side
impact than forward-facing restraints
that the transition to forward-facing
should be delayed as long as possible
(Henary et al. 2007). Injuries to the
head and face are most common in side
impacts, so restraints with larger padded
sidewings may offer some protection
(Orzechowski et al. 2003, Arbogast
et al. 2010, Maltese et al. 2007).
Laboratory testing of child restraints
with different types of LATCH hardware indicate that rigid LATCH
offers improved protection by limiting
motion towards the struck side of the
vehicle (Klinich et al. 2005). Tests
using a tether and forward-facing child
restraints showed a negligible effect
on lateral head excursion compared to
those without a tether (Klinich et al.
2005). Testing with additional energy
absorbing elements (side air cushion)
showed a significant improvement
over a baseline design (Bendjallal et al.
Forward-facing child restraints and
Although all children are safer in
the rear seat, if all of the rear seating
positions are occupied, a child in a
forward-facing harnessed child restraint
would usually be the best candidate to
ride in the right-front passenger position. A child well secured in a properly
installed forward-facing child restraint
should be at no greater risk of injury
from airbag deployment than a belted
adult in the same seating position.
The use of a harness around the child
reduces the likelihood of being outof-position and thus close to an airbag
than an older child in a booster seat or
seatbelt. Of the children sustaining fatal
injuries from deploying airbags, none
was seated in a properly used forwardfacing child restraint (NHTSA 2009). If
a forward-facing child restraint needs to
be installed in the right-front position,
the vehicle seat should be positioned
as far rearward as possible, while still
allowing for accommodation of the rear
seat occupant.
UMTRI Research Review
Figure 15. Without a booster (top), seatbelt loads the abdomen.
January - March 2012
How boosters improve belt fit
Over the past decade, evaluations
with child volunteers have examined
how different booster seat designs
improve belt fit using realistic vehicle
and seat belt geometries (Reed et al.
2008, 2009, Bilston and Sagar 2007).
This research has led to a better understanding of how booster seats improve
belt fit.
Recommended use and effectiveness
When a child no longer fits in a
harnessed restraint, the next step is a
belt-positioning booster seat used with
a vehicle lap-and-shoulder belt. As
with the transition from a rearward-to
forward-facing child restraint, this step
to a booster actually decreases the level
of occupant protection offered and
should be delayed as long as possible.
Boosters do not restrain children.
Instead, they reposition the child and
redirect vehicle belts (designed to fit
adults) to be routed appropriately
relative to the child’s body. Both the
NHTSA and the AAP recommend that
children use booster seats until they
fit in seat belts alone, which means
most children should be using boosters
through age 8–12 years (AAP 2011).
Booster seat use among 4-to-8 year
olds has risen to 63% in 2007 from 15%
in 2000, largely as a result of state laws
requiring their use, public education
programs, and more available booster
products (NHTSA 2009).
Children aged 4–8 using boosters are 45% less likely to sustain injury
in a crash compared to children
using seatbelts alone (Arbogast et al.
2009, Durbin et al. 2003). Boosters
are particularly effective at reducing abdominal injury: children using
belts alone are 8 times
more likely to sustain
abdomen injury than
children using a beltpositioning booster
with the vehicle seatbelt (Jermakian et al.
2007). Figure 15 illustrates consequences
from simulated frontal
crashes for a 6YO with
and without a booster.
With a booster, the lap
belt loads and restrains
the strong bones of
the pelvis. Without a
booster, the dummy
slides under the lap
belt, so the belt loads
the abdomen, vulnerable internal organs,
and spine instead of
the pelvis. This event is
often called “submarining” under the lap belt.
Figure 16. A 7YO child seated on the vehicle seat (top) has the
shoulder belt against the neck and the lap belt over the abdomen.
Use of a booster seat (bottom) shifts the child relative to the belts
so they fit to provide better protection.
the lap belt angle, pictured in figure
17. The steeper lap belt angle is better
because it makes it harder for the child
to slide under the lap belt in a crash.
The second way boosters work is
by improving occupant posture. Several
studies have documented that the rear
seats of most vehicles are too deep
for children to sit upright with their
knees bending over the edge of the
Figure 18. When seated on the vehicle seat, a 7YO child tends to
slouch forward (top). A booster seat (middle) lets the 7YO child
sit upright more comfortably and obtain a posture similar to that
achieved by a taller child on the vehicle seat (bottom).
Figure 17. Lap belt angle increases with booster use (even without
lap belt guides), allowing better engagement of the lap belt with
the pelvis.
seat and with their back fully supported
for comfort (Huang and Reed 2006,
Klinich et al. 1994, Bilston and Sagar
2007). Consequently, children scoot
forward so their legs can bend over the
front of the seat in a comfortable position, as shown in figure 18 (top). Using
a booster seat provides them with a
cushion length that is more compatible
with their upper leg length (figure 18
center) and provides an upright posture
similar to that of an older child (figure
18 bottom).
The third way boosters work is
by routing the seatbelt using lap-andshoulder belt guides. The lap belt
should be positioned so it is completely
below the top of the pelvis, which
reduces the likelihood that it will
slide up over the abdomen in a crash.
Well-designed lap belt guides help position the belt so it touches the top of
the child’s thighs, and resists upward
movement of the belt in a crash. Welldesigned shoulder belt guides position
the shoulder belt midway between the
neck and arm, not at the edge of the
shoulder or rubbing the neck (figure
19). Neck injury from the shoulder
belt contacting the neck has not been
identified as a problem in the field.
The biggest danger from the shoulder
belt touching the neck is that it could
cause the child to put the shoulder
belt under the arm or behind the back.
Either misuse virtually eliminates upperbody restraint that the properly placed
shoulder belt would provide. In one
study of booster misuse, 20% of children improperly placed the shoulder
belt behind the back or under the arm
(O’Neil et al. 2009). Poorly designed
shoulder belt guides can pull the shoulder belt too far off the child’s shoulder,
or allow slack to develop after a child
leans forward because it interferes with
easy retraction of the shoulder belt.
The first thing that any booster
seat does is raise the child up relative
to the vehicle belt as shown in figure
16. Even if the booster does not have
a back, the elevation helps position the
shoulder belt away from the neck so it
is more comfortable and restrains the
child through the shoulder structure in
a crash. When considering the lap belt,
shifting the child upwards relative to
where the lap belt is anchored increases
Figure 19. Shoulder belt guide positions the shoulder belt closer
to the child.
UMTRI Research Review
Built-in belt-positioning booster that stows in the vehicle
seat cushion.
January - March 2012
Over the years, several vehicle
manufacturers have offered the
option of built-in or integrated
child restraints in their vehicle
seats. While harnessed, forwardfacing, built-in restraints have
been produced in the past, today
only built-in booster seats that pop
up from the vehicle seat cushion
(Jakobssen et al. 2007) are
currently available as an option on
a few vehicle models (see below).
The advantage of a built-in child
restraint is that it links the child
directly to the vehicle and eliminates errors in installing the child
restraint to the vehicle. Arguments
against built-in harnessed restraints
are that rear-facing models have
not been offered, a child could
only use a harnessed forwardfacing restraint for up to four
years, and the restraint could not
be transferred for use in other
vehicles. A built-in booster may
prove to be more popular, as a
child could potentially use it for
four-to-eight years, and recent
commercial versions allow the
children to enjoy the comfort and
safety features of the vehicle seat
back. In addition, older children
who should use a booster may be
less likely to resist extended use if
it is part of the vehicle seat.
Changes in booster use and design
There are currently four styles
of belt-positioning boosters: backless
boosters, removable-back boosters,
highback boosters, (figure 20) and
built-in boosters. Backless boosters can
be used when the vehicle seat and head
restraint support the child’s head to the
tops of the ears. Some backless boosters
have an optional shoulder belt guide
on a strap to adjust the shoulder belt
position if necessary. With removableback boosters, the lower portion can
be used alone or with a booster seatback. Highback boosters are usually
constructed as combination seats that
can be converted from harnessed
restraints. A few vehicle manufacturers
provide integrated booster seats that
fold out or pop up from the vehicle
seat (Jakobssen et al. 2007).
Results from field data show that
there was no difference in injury risk
between boosters with and without
backs (Arbogast et al. 2009). While
boosters with backs have features
that could improve protection in side
impacts and may keep children in a
better position laterally relative to the
vehicle belt system particularly when
sleeping, backless boosters allow children to sit further rearward, which
effectively reduces head excursion.
From a practical standpoint, backless
boosters and built-in boosters allow
children to enjoy the comfort features
of a vehicle seatback, and since they
are not as visible from outside the
vehicle, they may be preferred by older
children reluctant to use a booster.
In addition, one study of children in
boosters showed that children seated
in products with large side wings for
improved side-impact protection leaned
forward 55% of the time compared to
25% of the time for children seated in
boosters with less prominent side wings
(Andersson et al. 2010).
Figure 20. Backless booster (top), removable back booster
(center) and high-back booster (bottom).
While many boosters with backs
have shoulder-belt positioning devices
that improve static belt fit, research
has indicated that the devices are not
that effective at keeping the shoulder
belt position in place during dynamic
loading (Tylko and Dalmotas 2005,
a child will not stay in position, some
have suggested locking the shoulder
belt with its switchable retractor (if
available). However, this does not allow
enough forward motion of the torso,
which prevents submarining under the
lap belt in the absence of a crotch strap.
A child who is not developmentally
ready to sit still in a booster would
be better protected in a high-weight
harness child restraint.
Klinich et al. 2008). Based on these
results, it is best to choose a booster
and vehicle seating position that
achieves good shoulder belt fit with
minimal redirection of the shoulder belt
by the booster. As shown in figure 21,
it would be better to have a straighter
line path between the D-ring and
shoulder (top) than one substantially
rerouted by the booster (bottom).
Booster seats must meet dynamic
FMVSS 213 requirements using a
test bench equipped with only one
defined lap-and-shoulder belt geometry. However, lap-and-shoulder belt
geometry in the rear seats of vehicles
can vary widely. Some boosters may not
be able to route belts with a particular
geometry so the belt will fit well on
a particular size of child. In practice,
the best approach is to evaluate the
belt fit with the specific child, vehicle
seating position and booster seat.
Several studies have documented that
the effectiveness of the booster seat
routing features varies with vehicle belt
geometry (McDougall 2011, Brown et
al. 2009). The Insurance Institute for
Highway Safety has developed a rating
system for assessing the belt fit across a
range of vehicle belt geometries (Reed
et al. 2009). However, given the effectiveness of booster seats demonstrated
in field data and the many factors that
allow boosters to improve belt fit, any
booster is likely to provide better seat
belt fit for a child than the no booster
Some children making a transition
from a harnessed restraint to a beltpositioning booster often have trouble
staying correctly positioned for the
entire trip, as the shoulder belt’s emergency locking retractor comfort features
allow considerable movement unless
activated and locked during a crash. If
Securing a booster in a vehicle
When first introduced, beltpositioning booster seats were not
secured to the vehicle, as their purpose
is to position a child relative to the
vehicle seat belt, but not to actually
provide restraint. However, the lack
of attachments sometimes allowed
the boosters and child to shift during
driving and caused instability during
loading/unloading. Since many
caregivers do not fasten the seatbelt
around the booster when unoccupied
as directed, a loose booster could be a
projectile in a crash.
There are some booster products
designed to allow the booster to be
secured to the lower anchorages and/
or tether anchorages with the LATCH
hardware. This is most common among
boosters that convert from a harnessed
restraint to a belt-positioning booster
and thus have LATCH attachments.
Some boosters also have rigid or flexible LATCH attachments solely to hold
them in place (SafeRideNews 2011).
This practice has not been universally
adopted, because there are lingering
concerns among some manufacturers
that if the booster and seatbelt but not
the child are attached to the vehicle,
the child could slide forward on the
booster and have a greater risk of
injury than if the booster moved with
the child. Testing results with boosters
attached to vehicles have been mixed,
with some tests showing improved
kinematics using a LATCH-secured
booster, and others showing less desirable kinematics (Tylko et al. 2005,
Transport Canada 2011).
Figure 21. A booster in a seating position with a relatively straight
path between the D-ring and shoulder (top) is preferred to one
requiring more redirection of the shoulder belt (bottom).
UMTRI Research Review
Shield boosters, originally
designed for use in seating positions with only a lap belt, are
no longer sold in the United
States. The greater availability
of harnessed restraints accommodating children of higher
weights, add-on harness systems,
and the requirement to have
lap-and-shoulder belts in all rear
seating positions, all provide
better options with superior torso
restraint than that of a shield
booster (Stalnaker 1993, Whitman
et al. 1997, Marriner et al. 1995,
Shelness and Jewett 1983).
Boosters and Lap Belts
Belt-positioning booster seats are
not designed to work with lap-only
belts, as they cannot pass FMVSS 213
head excursion requirements when used
this way. While using a booster seat
with a lap belt may reduce the likelihood of abdomen injury (Kirley et al.
2009), it has the potential to increase
the likelihood of head injury, which
should be considered higher priority
because of the greater potential for
serious long-term consequences. A
booster positions a child’s head higher,
and with a highback booster, more
forward than a child sitting directly on
a vehicle seat. Without torso restraint
provided by a shoulder belt, the head
position of a child using a booster
increases the risk of head contact
compared to a child on a vehicle seat.
With respect to prevention of head
contact, it is better for a child to sit
directly on the vehicle seat when only a
lap belt is available than to sit on a belt
positioning booster.
Seatbelts for Children
The term seatbelt refers to either a
lap-and-shoulder combination or a lap
belt alone. Although lap-and-shoulder
belts have become standard equipment in current vehicles and seating
positions, there are still many vehicles
on the road with only lap belts in rear
seats. Vehicle seatbelts are designed
primarily with adults in mind, and
geometric factors may make good fit
difficult for children. However, if a
more appropriate restraint system is
unavailable, seatbelts provide some
protection even for small children,
and effectiveness rates for seatbelts
are calculated for occupants age 5 and
up to be near 50% in terms of reducing fatal injuries (Wiacek et al. 2011).
Seatbelts are part of the continuum of
restraint systems with varying levels of
effectiveness for children. In general,
more restraint is better than less, and
good fit is important for effective
restraint performance. Unfortunately,
poor fit of seatbelts often leads to
misuse, with shoulder belts placed
behind the back or under the arm
(Louman-Gardiner 2008, Gotschall
et al. 1998, Meissner et al. 1994),
which degrades their performance and
increases the likelihood of submarining
and belt-induced injury.
Child size and belt fit
Belt fit depends on the size and
posture of the occupant, the size and
shape of the vehicle seat, and the
geometry and features of the belt
system. A child who has good belt fit in
one vehicle may not in another. Good
fit of a lap belt is as low as possible
on the pelvis, touching or flat across
the tops of the thighs. A child can
locate the top of his or her own pelvis
by finding the bony points at the top
front of the pelvis. A child’s pelvis is
generally shorter, less calcified and
less prominent than that of an adult.
January - March 2012
Therefore, it is critical that the lap
belt should lie completely below these
points to ensure that the lap belt can
contact and restrain the pelvis during
a crash (figure 22). The shoulder belt
should lie flat on the shoulder about
halfway between the neck and the arm
and cross the chest at the middle of the
A common recommendation is
that children should not use a seatbelt
without a booster until they reach a
standing height of 148 cm (58 in)
and a clothed weight of 37 kg (81
lb) based on an early study of booster
belt fit (Klinich et al. 1994). This
size corresponds to a 90th percentile
9-year-old, a 50th percentile 11-yearold, and a 5th percentile 13-year-old.
While a simple height, age or weight
recommendation is convenient for
educational or legislative purposes,
several studies indicate that most children above this stature still experience
better belt fit with a booster
To achieve the best seat belt
fit, the child should be sitting fully
upright with his/her pelvis as far
back into the seat as possible, and
preferably with his/her feet touching
the floor. This will help place the lap
belt in front of the pelvic bone below
the anterior-superior iliac spines and
will minimize the possibility of the
belt sliding up and intruding into the
soft upper abdomen. Several studies
have shown that children tend to
move forward on the vehicle seat to
allow their knees to bend comfortably over the front edge of the seat,
causing the child to slouch. This
rotates the pelvis rearward, making
it more difficult for the lap belt to
engage the pelvis, and can lead to
the lap belt being positioned over
the abdomen. If a child cannot
achieve an upright, seated posture,
or if the shoulder belt crosses the
throat, the child needs to use a
Shoulder belts that touch the
side of the neck are not likely to
cause injury unless they are very
loose (Kortchinsky et al. 2008,
Corben and Herbert 1981, Appleton
1983). However, discomfort from a
shoulder belt against the neck tends
to cause the child to put the shoul-
der belt under his or her outboard
arm or behind the back. The shoulder belt should not be routed behind
the child’s back because it offers
no torso restraint and tends to pull
the lap portion of the belt upwards
on the inboard side, both of which
increase potential for injurious belt
loading (Brown and Bilston 2007).
Also, with most belt designs, routing
behind the back eliminates the
loading of the shoulder belt early
in the crash sequence, which, on a
properly worn belt, functions to snug
the lap belt and, in some retractor
designs, lock the belt. Finally, the
shoulder belt should never be routed
under the arm, because the resulting
belt forces on the side of the thorax
are known to result in serious internal injuries in a crash (Gotschall et
al. 1998, States et al. 1987).
It is possible for shoulder belt
loading to cause thorax injury in
severe crashes as it loads the child.
To reduce likelihood of injury from
belt loading, advanced seat belt
features which have been implemented for front-row occupants are
gradually being introduced in the
rear seating positions. One of these
features is a pretensioner, which
removes slack from the seatbelt when
a crash event is detected. Another
feature is a belt load limiter, which
allows the shoulder belt to spool out
further once a particular load threshold is reached. An airbag mounted in
the shoulder belt to provide better
load distribution over the thorax has
been introduced on rear-seat belts in
some vehicles.
Vehicle manufacturers have also
added seatbelt features to improve fit
for various sizes of occupants. Many
vehicles have an adjustable shoulder
belt anchorage that can be raised
Figure 22. On a 7YO child seated on a vehicle seat (top left), the vehicle belt lies on the abdomen above the pelvis. When using a booster or lowered to better route the belt
(top right), the lap belt lies below the top of the pelvis. For a 10YO child on the vehicle seat, the lap belt lies just below the pelvis (lower
over the occupant’s shoulder. Some
left), providing acceptable fit. Even so, the 10YO using the booster experiences much better fit over the hips (lower right).
UMTRI Research Review
have positioning guides or loops that
can also help provide better fit for
smaller occupants. However, these
may not help fit problems with lap
belts or vehicle seat cushions that are
too long for a child to sit upright
Lap versus lap-and-shoulder belts
Fortunately, the relatively recent
requirement to provide lap-andshoulder belts in all vehicle rear
seating positions has reduced the
need to use only lap belts to restrain
occupants. The principles of restraint
theory lead to the conclusion that
lap-and-shoulder belts would be
January - March 2012
Lap-and-shoulder belts and airbags
Even with advanced airbags,
which can sense and adjust deployment to the size and type of front
passenger, parents are warned not
to have children under age 13 ride
in the front seat of a vehicle. Older
vehicles such as small pickup trucks
without advanced airbags may have
on/off switches for frontal airbags.
In situations where a child must
ride in the front seat with an active
airbag, because no switch is available
and the back seat is filled, a child in
a seatbelt may be at greater risk of
injury from the frontal airbag than
a younger sibling restrained in a
forward-facing child restraint. This is
because the child using the lap-andshoulder belt is able to lean forward
in their shoulder belt or even put the
belt behind the back. This behavior
may place the child’s head in the path
Shoulder belt positioners
Various unregulated devices have
been marketed to move a shoulder
belt away from a smaller occupant’s
neck. Most of these products pull
the shoulder belt into position by
anchoring a device to the lap belt,
thereby pulling that portion of the
belt upward and gaining shoulder
belt fit at the expense of proper lap
belt fit (Brown et al. 2010, Sullivan
and Chambers 1994). Unlike a beltpositioning booster, shoulder belt
positioners typically pull the lap belt
up onto the abdomen as they pull
the shoulder belt down and away
from the neck. In addition, they do
nothing to improve the posture and
slouching of a child too small to fit
in the vehicle seat. Because pediatric
dummies cannot currently measure
loading to the abdomen, evaluation
of the potential negative effects of
shoulder belt positioners cannot be
quantified. These products may be
packaged with misleading claims that
they “meet all relevant standards”
when none apply. Shoulder belt positioners should not be used in place of
belt-positioning boosters, which are
proven in the field to reduce injury,
particularly to the abdomen.
better for children, even if fit is
not optimal, than a lap belt alone.
Analysis of fatality data confirms that
lap-and-shoulder belts are 15% more
effective than lap belts alone. While
lap-only belts reduce the risk of ejection and injury, they increase the risk
of abdominal injuries; lap-and-shoulder belts reduce the risk of both head
and abdominal injuries (Elliott et al.
2006, Mulpuri et al. 2007, Morgan
of the deploying airbag or allow their
upper body to be thrown forward
during precrash braking.
Test Procedures
FMVSS 213 Testing
Given the high level of occupant protection provided by current
child restraints in all types of crashes,
people are often surprised to find
that the testing requirements as
defined in the applicable federal
rule, FMVSS 213, primarily focus on
their performance in frontal crashes
at one severity level (CFR, FMVSS
213). In addition, though vehicle
seats, LATCH anchorage locations,
and seatbelt geometry vary widely in
vehicles, child restraints are tested on
a generic, soft, flat, bench seat using
either a single set of belt anchorages or LATCH anchorages. Child
restraints are not crash tested in real
vehicles, but tested using a sled that
simulates the acceleration seen in a
crash with a 30 mi/hr (48 km/hr)
change in velocity. Sled tests are used
because they are more repeatable
and less expensive. At first glance, a
30-mi/hr (48 km/hr) test may not
seem very severe, but 30 mi/hr (48
km/hr) refers to the change in velocity, not the velocity at the time of the
crash, and the crash conditions used
are more severe than 96% of actual
frontal crashes in the United States.
When evaluating the dynamic
safety performance of a child
restraint, requirements vary with the
type of restraint. For a car bed, the
primary criterion is that the harness
must keep the newborn size dummy
in the restraint. For rear-facing
restraints, a restraint will pass if the
surface supporting the crash dummy’s
back does not rotate forward beyond
an angle of 70°, the head and chest
of the dummy stay in the restraint,
and the acceleration characteristics
for the dummy’s head and chest do
not exceed prescribed thresholds.
For forward-facing restraints and
booster seats, the dummy’s head
must not move forward past a point
720 mm (28.4 inches) from a seat
reference point when tested with a
tether or 813 mm (32 inches) when
tested without a tether; in either
condition, the knees must not pass
a point 915 mm (36 inches) away.
In addition, there are limits on
head and chest acceleration-based
measures. Unfortunately, better
scores on the head injury criteria
can usually be achieved by allowing more head excursion, although
keeping the head from moving
further forward corresponds to
preventing the most common realworld head injury mechanism of the
child’s head striking something in
the vehicle (Bohman et al. 2011).
Other 213 requirements focus on
webbing strength, width, and abrasion resistance, flammability of the
components, buckle release characteristics, and padding requirements.
Tests are also performed to determine
whether the child would stay within
the restraint when it is inverted.
Requirements for LATCH hardware in vehicles are specified in
FMVSS 225 (CFR FMVSS 225).
Most vehicles have the minimum
required LATCH hardware where
top tether anchorages are provided
in three seating positions and
lower anchorages are provided in
two seating positions. The regulation defines zones for locating the
lower and tether anchorages, as well
as quasi-static (or slow loading)
testing procedures for evaluating the
strength of the lower and top tether
anchorages. Other requirements for
lower anchorages include specifications for the size and spacing of the
anchor bars that comprise the lower
anchorages, and requirements for
how a child restraint fixture must fit
in the vehicle when attached to the
lower anchorages. The lower anchorages must either be visible or labeled,
but there are no labeling requirements for tethers.
Side impact testing
Vehicle-to-vehicle side impact
events are often described based on
the occupant’s position relative to the
striking vehicle. If the occupant is on
the opposite side of the vehicle from
the striking object it is called the
“far-side” impact condition and a seat
belt can play an important role in the
outcome by limiting the possibility
of occupant contact. When the occupant is positioned on the side of the
vehicle closest to the striking vehicle
it is called a “near-side” impact event
and injuries are often caused by direct
loading between the striking object
and the occupant. In near-side events,
use of the seat belt is less of a factor
in the crash injury outcomes.
Near-side impacts are most injurious, and the occupant motions
involve the child restraint moving
toward the door as the door is
intruding from the striking vehicle.
US child restraint products currently
do not have to be tested under side
impact loading conditions. However,
many child restraint manufacturers
advertise that they have tested their
products in side impact using internal
test procedures. Side impact tests are
generally conducted with dummies
that are designed for side impact
evaluation. In addition, the simulated
side impact crash is run at a lower
change in velocity than frontal impact
testing to reflect the typical crash
severity in the field.
Many different strategies have
been proposed for testing child
restraints in side impact to approximate the loading conditions seen in a
vehicle. Child restraint manufacturers
likely use some combination of these
tests. Methods include:
•Repositioning the bench used for
frontal impact testing and decelerating the child restraint laterally.
This represents the loading that
a child restraint would undergo
in a far-side crash. This type of
testing does not represent the
most injurious side impact loading,
but can demonstrate how well the
attachment system keeps the child
restraint from moving laterally
and how well the dummy’s head is
contained within the restraint.
•Lateral loading into a fixed rigid
wall. The main difference between
this method and the previous one
is that the test fixture includes a
rigid plate mounted at the end
of the seating bench. This testing
method is used in Australian
regulations (AS/NZS 1754). In
addition to demonstrating the
ability to prevent lateral movement
and contain the dummy’s head,
this method allows a rough assessment of head injury potential from
contacting a vehicle surface.
•Lateral loading into a rotating
door. This approach, considered
for European testing, was thought
to provide a way of approximating
intrusion. The characteristics of
the door have a significant effect
on the loading. It was difficult
to achieve consensus on what the
door characteristics should be as
the design of vehicles has changed
over time in response to vehicle
side-impact requirements.
•Simulated door intrusion. This
strategy propels a simulated door
into the side of a fixed child
UMTRI Research Review
January - March 2012
injury risk to adjacent occupants. In
addition, testing procedures evaluate injury risk by measuring lateral
head excursion, and most rear-facing
restraints have higher values than
forward-facing seats even though
they are demonstrated to be safer in
crashes. Comparison of values may
encourage caregivers to inappropriately shift to forward-facing restraints
Vehicle testing
In addition to testing of child
restraints, vehicles must meet regulatory requirements that pertain to
protection of child occupants. Vehicle
manufacturers perform a series of
tests to ensure that frontal airbags do
not deploy at injurious levels when
a child occupant is in the right-front
seating position, including when
they are “out-of-position” and close
to the airbag module (CFR FMVSS
208). Vehicle manufacturers also
perform voluntary testing to check
that side airbags do not pose a danger
to children (Side Airbag OOP IT
Working Group 2003). FMVSS 201
defines tests for evaluating the injury
potential if occupants contact interior structures of the vehicle, such as
the roof and B-pillars (CFR FMVSS
201). While children benefit from the
interior padding and energy-absorbing structures that result from these
requirements, the requirements do
not apply to many of the structures
in the rear seat that are commonly
contacted by child occupants during
crashes because the regulation
primarily addresses interior points at
or above the window sill (Arbogast
et al. (in press), Jermakian et al.
2007). FMVSS 214, which evaluates
the safety of vehicles in side impacts
using adult-sized crash dummies,
also benefits child occupants (CFR
FMVSS 214).
restraint. This approach captures
most of the kinematics of a nearside crash except for the initial
movement of the child restraint
towards the intruding door. An
example of this approach is the
new European Union regulation
that uses a moving sled to propel
the child restraint into a padded
fixed door and the side impact test
fixture developed by Dorel and
Kettering University.
•Simulated door intrusion including child restraint motion. Takata
Corporation developed a side
impact sled test method that simulates the door intrusion typical of
a crash with a near-side occupant
in a child restraint system. The
method employs a base structure
that simulates the vehicle door and
a separate vehicle seat that slides
on rails and moves into the door
structure during the crash event.
Honeycomb aluminum is positioned between the two elements
to simulate the crush of the vehicle
structure. The door structure is
padded to simulate the compliance
of a vehicle door. The method
can be used to run a pure side
impact or a side impact crash with
a frontal deceleration component
by adjusting the mounting angle
of the entire buck relative to the
primary direction of sled movement. (Sullivan and Louden 2009,
Sullivan et al. 2011)
While the idea of testing child
restraints in side impact has merit,
design changes in response to side
impact testing may have unintended
consequences. If child restraints
become wider to accommodate
padding or larger sidewings, they
may be more difficult to install with a
child restraint in an adjacent seating
position. Restraints may also become
heavier and stiffer, possibly posing an
Child restraints are not currently
tested in vehicle crash tests for regulatory purposes. However, vehicle
designs developed to improve safety
for adult occupants in regulatory and
consumer testing may benefit child
occupants as well. Research has been
conducted using child restraints and
pediatric dummies in a number of
test programs that have identified
possible issues with child restraint
performance in severe crashes (Park
et al. 2011, Tylko 2011). These
results have led to additional research
programs to identify methods of
improving the safety of the rear
seating compartment for the child
occupants who primarily sit there
(Hu et al. 2011, Klinich et al. 2008
and 2011, Reed et al. 2008).
Injury Criteria Limitations
The measure of head injury
potential traditionally used in dummy
testing is called HIC (head injury
criteria). HIC involves integrating
the measured head accelerations over
a particular time period, and was
originally developed to correspond
with likelihood of skull fracture
from direct loading (Versace 1971).
Dummy Limitations
Pediatric crash dummies are
designed to have the dimensions
of an average child of the age they
represent, primarily based on a 1977
study of child anthropometry (Snyder
et al. 1977). Their overall weight
matches the average child weight
from this study as well, but the
distribution of weight among body
segments is scaled from the distribution found in adults. The response
to loading of child dummies is also
scaled down from adult dummies
with limited adjustment made for
changes in mass and stiffness (Irwin
and Mertz 1997). The responses of
adult dummies are primarily based
on testing of elderly cadavers under
dynamic loading conditions. Most
pediatric dummies are designed for
frontal impact loading, although
some versions are designed for side
impact testing.
Results from testing using crash
dummies must be viewed within the
limitations of the data from which
they were developed. On the one
hand, child crash dummies have been
used to develop the safest restraints
available. On the other hand, child
dummies do not sit the way children
do, have limited amounts of sensors/
instrumentation, and have idiosyncrasies that can affect test results (Ash
et al. 2009). When neck loads have
been measured in child dummies
seated in restraint systems, they often
reach alarming levels relative to estimates of neck injury thresholds based
on scaling adult values (Park et al.
2011, Menon et al. 2005). Given
that serious neck injury to properly restrained children is rare, neck
injury-related measurements need
to be reviewed with caution. The
Injury threshold values for HIC were
scaled for children from adult data
(Irwin and Mertz 1997). While HIC
seems to work reasonably to predict
head injury from head strike, high
HIC values can also arise from the
dummy’s head moving rapidly during
deceleration without contacting
Since the main pediatric head
injury mechanism of direct contact
with vehicle interior components
is not simulated with the FMVSS
213 test fixture, the use of HIC as a
measure during FMVSS 213 testing
may be somewhat flawed. Head
excursion, which is also evaluated
during FMVSS 213 testing, is likely a
better predictor of head injury potential, in that the further forward the
head travels during loading, the more
likely it will strike a vehicle interior
component (even if that vehicle interior is not represented on the FMVSS
213 test buck.) Head injury from
vehicle interior contact is the most
common mechanism of pediatric
head injury in crashes (Bohman et al.
2011). Nance et al. (2010) studied
factors associated with clinically
significant head injury and their findings for impact type and vehicle size
suggest head contact as a mechanism.
FMVSS 213 also places limits on
the allowable thoracic loading based
on the measured chest acceleration.
However, serious chest injuries in the
absence of significant intrusion are
also relatively rare in field data.
design of the child dummy’s spine
only has flexible components in the
neck and lumbar, not the thoracic
region, which may be leading to
more bending and higher loading
in the neck than a real child with a
fully flexible spine would experience
(Seacrist et al. 2010).
Another example is that standard child dummies cannot currently
measure abdominal loading, which
is one of the more common body
regions injured in older children not
using boosters. The movement of
the dummy during dynamic testing
must be reviewed as well as the values
from instrumentation, and both
must consider the limitations of the
dummy and instrumentation. Design
of good child restraints must balance
test results with field data and judgment.
UMTRI Research Review
The consistent and proper
use of restraint systems by infants
and children in passenger vehicles
prevents hundreds of deaths and
thousands of injuries each year.
Misuse or improper selection of
child restraints as well as nonuse by
a small minority of children leads
to many of the fatalities that do
occur. Infants require the highest
level of special treatment, with
restraint systems designed to apply
crash forces along the full length
of their bodies. Toddlers can also
benefit from rear-facing restraints.
All children are best protected by
harnessed restraints that snugly
conform to their small body shape
and are tightly installed in the
vehicle. Belt-positioning boosters improve posture and belt fit so
the vehicle seat belts can effectively
protect older children in crashes.
Seatbelts can provide good protection for children approaching the
size of adults if the lap belt fits so
it loads the pelvis and the shoulder belt fits so it loads the clavicle.
Understanding both the theory
behind the design of restraint
systems and its application to child
restraints is needed to develop
improved restraint systems, and to
provide informed guidance concerning child restraint selection and use.
We would like to thank Britax
Inc. for financial support of this
effort. We would also like to thank
the following experts for reviewing
this document.
January - March 2012
Kristy Arbogast, PhD
Alisa Baer, MD
Farid Bendjallal, PhD
Denise Donaldson
Shashi Kuppa, PhD
Matt Reed, PhD
Jonathan Rupp, PhD
Lawrence Schneider, PhD
Deborah Stewart
Lisa Sullivan
Stephanie Tombrello
Learn more on our website at
Dr. Kathleen D. Klinich
is an assistant research scientist
at the University of Michigan
Transportation Research Institute.
Her research interests focus on
protecting occupants in crashes,
particularly child passengers.
Research projects include pediatric dummy development, analysis
of crash datasets, factors affecting
child restraint and vehicle compatibility, investigations of belt fit on
child occupants, analysis of crashes
involving airbags and rear-facing
infants, and testing to evaluate
pediatric crash dummy performance under conditions other than
FMVSS 213. She is a member of the
SAE Children’s Restraint Systems
Standards and Dummy Testing
Committees. She is a NHTSA certified child passenger safety technician
and mother of three.
Miriam A. Manary is a
senior engineering research associate at the University of Michigan
Transportation Research Institute
and has been a principal investigator on studies focused on child
passenger safety and wheelchair
transportation safety. She supervises
the child restraint testing program at
UMTRI including testing to FMVSS
213 standards and product development programs. As a member of
the ISO child restraint standards
committee, she has been involved
in the development of worldwide
child passenger safety standards. Ms.
Manary is also the cochair of the
SAE Children’s Restraint Systems
Standards Committee and serves
on the editorial board of Safe Ride
News. She is a NHTSA certified
child passenger safety instructor and
a mother of three.
Kathleen B. Weber is an emeritus researcher from the University of
Michigan who authored the original document, “Crash Protection
for Child Passengers: A Review of
Best Practice” (2000) www.umtri.
She was a pioneering researcher in
the field of child occupant protection, published numerous papers
on her research, and served on
SAE, ISO, and advisory panels that
provided the foundation for current
best practices in child passenger
American Academy of Pediatrics (2011a). Car Safety
Seats: Information for Families for 2011.
American Academy of Pediatrics, Committee on Injury,
Violence, and Poison Prevention. (2011b). Policy
Statement: Child passenger safety. Pediatrics 127;788793.
American Academy of Pediatrics, Committee on Injury,
Violence, and Poison Prevention. (2011c). Technical
Report: Child passenger safety. Pediatrics 127;10601068.
American Academy of Pediatrics, Committee on Injury
and Poison Prevention. (1999a). Transporting children
with special health care needs. Pediatrics 104:988-992.
American Academy of Pediatrics, Committee on Injury
and Poison Prevention. (1999b). Safe transportation
of newborns at hospital discharge. Pediatrics 104:986987.
Andersson M, Bohman K, Osvalder AL (2010). Effect of
booster seat design on children’s choice of seating
positions during naturalistic riding. Ann Adv Automo
Med 54:171-80.
Appleton I. (1983). Young children and adult seat belts:
Is it a good idea to put children in adult belts? New
Zealand Ministry of Transport, Road Transport
Division, Wellington.
Arbogast KB, Durbin DR, Kallan MJ, Menon RA,
Lincoln AE, Winston FK.(2002). The role of restraint
and seat position in pediatric facial fractures. J Trauma.
Vol 52:693-698.
Arbogast KB, Durbin DR, Cornejo RA, Kallan MJ,
Winston FK. (2004). An evaluation of the effectiveness
of forward facing child restraint systems. Accid Anal
Prev 36(4):585-9.
Arbogast KB, Jermakian JS, Kallan MF, and Durbin DR.
(2009). Effectiveness of Belt Positioning Booster Seats:
An Updated Assessment Pediatrics 124:1281-1286
Arbogast KB, Kallan MJ. (2007). The exposure of
children to deploying side airbags: an initial field
assessment. Annu Proc Assoc Adv Automot Med.
Arbogast KB, Locey CM, Zonfrillo MR, Maltese MR.
(2010). Protection of Children Restrained in Child
Safety Seats in Side Impact Crashes. Journal of
Trauma. Vol 69(4) . Oct:913-923.
Arbogast KB, Wozniak S, Locey CM, Maltese MR,
Zonfrillo MR (in press). Head injury contact points for
restrained child occupants. Traffic Inj Prev.
Ash J, Abdelilah Y, Crandall J, Parent D, Sherwood C,
Kallieris D. (2009). Comparison of anthropomorphic test dummies with a pediatric cadaver restrained
by a three-point belt in frontal sled tests. ESV Paper
Bahlmann N, Plein A, Schnottale B, Kromeyer-Hauschild
K. (2009). Protection of overweight children in child
restraint systems. ESV Paper Number 09-0414
Bendjellal F, Sciclune G, Frank R, Grospietsch M,
Whiteway A, Flood W, Marsilio, W. (2011). Applying
side impact cushion technology to child restraint
systems. ESV Paper No.11-0138-O
Berg MD, Cook L, Corneli HM, Vernon DD, Dean JM.
(2000). Effect of seating position and restraint use on
injuries to children in motor vehicle crashes. Pediatrics
Bilston LE, Sagar N. (2007). Geometry of rear seats and child restraints compared to child anthropometry.
Stapp Car Crash J 51:275-98.
Bilston LE, Yuen M, Brown J. (2007). Reconstruction of
crashes involving injured child occupants: The risk of
serious injuries associated with sub-optimal restraint
use may be reduced by better controlling occupant
kinematics, Traffic Injury Prevention, 8:1, 47-61
Bohman, Arbogast KB, Bostro O. (2011). Head injury
causation scenarios for belted, rear-seated children in
frontal impacts. Traffic Injury Prevention, 12:62–70
DOI: 10.1080/15389588.2010.526159
Braver ER, Ferguson, SA, Greene MA, Lund AK. (1997).
Reductions in deaths in frontal crashes among right
front passengers in vehicles equipped with passenger
airbags. Journal of the American Medical Association
Braver ER, Scerbo M, Kufera JA, Alexander MT, Volpini
K, Lloyd JP. (2008). Deaths among drivers and rightfront passengers in frontal collisions: Redesigned
airbags relative to first-generation airbags, Traffic
Injury Prevention 9(1):48-58.
Braver ER, Whitfield R, Ferguson, SA. (1998). Seating
positions and children’s risk of dying in motor vehicle
crashes. Injury Prevention 4:181-187.
Brown J, Kelly P, Suratno B, Paine M, Griffiths M.
(2009). The need for enhanced protocols for assessing
the dynamic performance of booster seats in frontal
impacts. Traffic Injury Prevention, 10(1):58-69.
Brown J and Bilston L. (2007). Spinal injuries in rear
seated child occupants aged 8-16 years. ESV Paper No
Brown J, Kelly P, Griffiths M, Tong S, Pak R, Gibson
T. (1995). The performance of tethered and untethered forward facing child restraints. International
IRCOBI Conference on the Biomechanics of Impact.
IRCOBI, Bron, France, 61-74.
Bull MF, Engle WA (2009). The Committee on Injury,
Violence, and Poison Prevention and the Committee
on Fetus and Newborn Safe Transportation of Preterm
and Low Birth Weight Infants at Hospital Discharge
Pediatrics 123(5):1424-1429 (doi:10.1542/peds.20090559)
UMTRI Research Review
Bull MJ, Weber KB, Talty J, and Manary, MA (2001).
Crash Protection for Children in Ambulances.
45th Annual Proceedings of the Association for the
Advancement of Automotive Medicine. AAAM, Des
Plaines, IL.
Code of Federal Regulations (2011). Title 49,
Transportation, Part 571.213; Child restraint systems.
National Archives and Records Service, Office of the
Federal Register Washington, D.C.
Code of Federal Regulations (2006). Title 49,
Transportation, Part 571.225; Child restraint anchorage systems. National Archives and Records Service,
Office of the Federal Register Washington, D.C.
Code of Federal Regulations (2007). Title 49,
Transportation, Part 571.208; Occupant crash protection. National Archives and Records Service, Office of
the Federal Register Washington, D.C.
Code of Federal Regulations (2003). Title 49,
Transportation, Part 571.201; Occupant protection in
interior impact. National Archives and Records Service,
Office of the Federal Register Washington, D.C.
Code of Federal Regulations (2010). Title 49,
Transportation, Part 571.214; Side impact protection.
National Archives and Records Service, Office of the
Federal Register Washington, D.C.
Corben CW, Herbert DC. (1981). Children wearing
approved restraints and adult’s belts in crashes. New
South Wales, Traffic Accident Research Unit, Sydney.
Decina LE, Lococo KH, (2005). Child restraint system
use and misuse in six states. Accident Analysis and
Prevention 37:583-590.
Decina LE, Lococo KH, (2007). Observed LATCH use
and misuse characteristics of child restraint systems in
seven states. Journal of Safety Research 38:271-281.
Degrazia M, Guo CY, Wilkinson AA, Rhein L. (2010).
Weight and age as predictors for passing the infant car
seat challenge. Pediatrics 125(3):526-31. Epub 2010
Feb 8.
Durbin DR, Chen I, Smith R, Elliott MR Winston
FK (2005). Effects of seating positon and appropriate
restraint use on the risk of injury to children in motor
vehicle crashes. Pediatrics 115(3):e305-9.
Dukehart JG, Walker L, Lococo K, Decina LE, Staplin
L (2007). Safekids Checkup Events: A National Study.
SafeKids Worldwide, Washington DC.
Durbin DR, Elliott MR, Winston FK. (2003). Belt
Positioning Booster Seats and Reduction in Risk of
Injury among Children in Vehicle Crashes. JAMA
289(21): 2835-2840.
Eby DW, Kostyniuk LP (1999). A statewide analysis of
child safety seat use and misuse in Michigan. Accident
Analysis and Prevention 31:555-566.
Edwards J, Sullivan K. (1997). Where are all the
January - March 2012
children seated and when are they restrained? SAE
971550. Reprinted in Child Occupant Protection
2nd Symposium. Society of Automotive Engineers,
Warrendale, PA, 35-42.
Elliott MR, Kallan MJ, Durbin DR, Winston FK. (2006a).
Effectiveness of child safety seats vs seat belts in reducing risk for death in children in passenger vehicle
crashes. Arch Pediatr Adolesc Med. 160(6):617-21.
Elliott MR, Arbogast KA, Durbin DR (2006b). A latent
class analysis of injury patterns among rear-seated, seatbelted children. J Trauma; 61:1244 –1248.
Evans SL, Nance ML, Arbogast KB, Elliott MR, and
Winston FK. (2009). Passenger compartment intrusion as a predictor of significant injury for children in
motor vehicle crashes. J Trauma. 66:504 –507
Feles N. (1970). Design and development of the General
Motors infant safety carrier. SAE 700042. Society of
Automotive Engineers, New York.
Fitzharris MP, Charlton J, Bohensky M, Koppel S, Fildes
B (2008). Booster seat use by children aged
4-11 years: evidence of the need to revise current
Australasian standards to accommodate overweight
children. Medical Journal of Australia 188:328-331.
Ford Motor Company (2011). Vehicle Owner’s Manuals
Model Year 2011, Ford Motor Company, Dearborn,
Fuchs S, Barthel MJ, Flannery AM, Christoffel KK.
(1989). Cervical spine fractures sustained by young
children in forward-facing car seats. Pediatrics 84:348354.
Gotschall CS, Better AI, Bulas D, Eichelberger MR, Bents
F, Warner M. (1998). Injuries to children restrained in
2- and 3-point belts. Association for the Advancement
of Automotive Medicine 42nd Conference. AAAM, Des
Plaines, IL, 29-43.
Greenspan AI, Dellinger AM, Chen J. (2010). Restraint
use and seating position among children less than 13
years of age: Is it still a problem? Journal of Safety
Research 41:183–185
Hallman JJ, Brasel KJ, Yoganandan N, Pintar FA. (2009).
Splenic trauma as an adverse effect of torso-protecting
side airbags: biomechanical and case evidence. Ann
Adv Automot Med. 53:13-24.
Haricharan RN, Griffin RL, Barnhart DC, Harmon DM,
McGwin G. (2009). Injury patterns among obese children involved in motor vehicle collisions. Journal of
Pediatric Surgery 44: 1218–1222.
Henary B, Sherwood CP, Crandall JR, Kent RW, Vaca
FE, Arbogast KB, Bull, MJ. (2007). Car safety seats
for children: rear facing for best protection. Injury
Prevention;13:398–402. doi: 10.1136/ip.2006.015115
Hoy GA, Cole WG. (1993). The pediatric cervical seat
belt syndrome. Injury 24:297-299.
Hu J, Klinich KD, Reed MP, Kokkolaras M, Rupp JD.
(2011). Development and Validation of a Modified
Hybrid-III Six-Year-Old Dummy Model for Simulating
Submarining in Motor-Vehicle Crashes. Medical
Engineering and Physics doi:10.1016/j.medengphy.2011.08.013
Huang S and Reed M. (2006). Comparison of Child Body
Dimensions with Rear Seat Geometry. SAE Technical
Paper 2006-01-1142, 2006, doi:10.4271/2006-011142
Huelke DF, Mackay GM, Morris A, Bradford M. (1992).
Car crashes and non-head impact cervical spine injuries in infants and children. SAE 920563. Society of
Automotive Engineers, Warrendale, PA.
Insurance Institute for Highway Safety (2003). Status
Report Volume 38 Number 5, June 11, 2003.
International Standards Organization (1999a). Road
vehicles - Anchorages in vehicles and attachments to
anchorages for child restraint systems - Part 1: Seat bight
anchorages and attachments. ISO/IS 13216-1. ISO,
Geneva, December 1999.
International Standards Organization (1999b). Road
vehicles - Child restraint systems - Reduction of misuse
risk - Part 2: Requirements and testing procedure for
correct installation (panel method). ISO/IS 13215-2.
ISO, Geneva, November 1999.
International Standards Organization (1999c). Road
vehicles - Child restraint systems - Reduction of misuse
risk - Part 3: Prediction and assessment of misuse by
Misuse Mode and Effect Analysis (MMEA). ISO/IS
13215- 3. ISO, Geneva, April 1999.
International Standards Organization (2008). Road
Vehicles – Methods and Criteria for Usability Evaluation
of Child Restraint Systems and Their Interface With
Vehicle Anchorage Systems – Part 1 Vehicles and Child
Restraints Equipped with ISOFIX Anchorages and
Attachments. CD 29061-1, Geneva, Switzerland.
International Standards Organization (2010). ISO
29061-1:2010 Road vehicles -- Methods and criteria
for usability evaluation of child restraint systems and
their interface with vehicle anchorage systems -- Part
1: Vehicles and child restraint systems equipped with
ISOFIX anchorages and attachments. ISO, Geneva,
Irwin A and Mertz HJ. (1997). Biomechanical basis
for the CRABI and Hybrid III child dummies. SAE
Technical Paper 973317. In Child Occupant Protection
2nd Symposium Proceedings (P-316), 261-274.
Warrendale, PA: Society of Automotive Engineers, Inc.
Isaksson-Hellman I, Jakobsson L, Gustafsson C, Norin H.
(1997). Trends and effects of child restraint systems
based on Volvo’s Swedish accident database. SAE
973299. Child Occupant Protection 2nd Symposium.
Society of Automotive Engineers, Warrendale, PA, pp
43-54, 1997.
Jakobsson L, Wiberg H, Isaksson-Hellman I, Gustafsson,
J. (2007). Rear seat safety for the growing child––a
new 2-state integrated booster cushion. ESV Paper
Number 07-0322
Jermakian JS, Kallan MJ, Arbogast KB. (2007).
Abdominal injury risk for children seated in belt positioning booster seats. ESV Paper Number 07-0441
Jermakian JS, Locey CM, Haughey LJ, Arbogast
KB (2007). Lower extremity injuries in children
seated in forward facing child restraint systems.
Traffic Injury Prevention, 8:171–179, DOI:
Jermakian JS, Wells JK. (2010). Observed use of tethers
in forward-facing child restraint systems. Inj Prev.
17(6): 371-4.
Johnson DL, Falci S. The diagnosis and treatment of
pediatric lumbar spine injuries caused by rear seat lap
belts. Neurosurgery 26:434-441 (1990).
Kahane CJ. An evaluation of child passenger safety—
the effectiveness and benefits of safety seats. DOT
HS 806 890. National Highway Traffic Safety
Administration, Washington, DC, February 1986.
Kallan MJ, Durbin DR, Arbogast, KB. (2008). Seating
patterns and corresponding risk of injury among
0-to-3-year-old children in child safety seats. Pediatrics
121:e1342 DOI: 10.1542/peds.2007-1512
Kapoor T, Altenhof W, Snowdon A, Howard A, Rasico J,
Zhu F, Baggio D (2011). A numerical investigation
into the effect of CRS misuse on the injury potential of
children in frontal and side impact crashes. Accid Anal
Prev 43(4):1438-50. Epub 2011 Mar 16.
King AI. (1993). Injury to the thoraco-lumbar spine and
pelvis. Accidental Injury: Biomechanics and Prevention.
Springer-Verlag, New York, pp 429-459.
Kirley BB, Teoh ER, Lund AK, Arbogast KB, Kallan
MJ, Durbin DR. (2009). Making the most of the
worst-case scenario: should belt-positioning booster
seats be used in lap-belt-only seating positions? Traffic
Inj Prev. 10(6):580-3.
Klinich KD, Pritz HB, Beebe MS, Welty K, Burton RW.
(1994). Study of older child restraint/ booster seat
fit and NASS injury analysis. DOT/HS 808 248.
National Highway Traffic Safety Administration,
Vehicle Research and Test Center, East Liberty, OH.
UMTRI Research Review
Klinich KD, Hulbert GM, and Schneider LW. (2002).
Estimating infant head injury criteria and impact
response using crash reconstruction and finite element
modeling. Technical Paper 2002-22-0009. Stapp Car
Crash Journal: 46:165-194.
Klinich KD, Ritchie NL, Manary MA, Reed MP,
Tambourra N, and Schneider LW. (2005). Kinematics
of the Q3S ATD in a Child Restraint under Far-side
Impact Loading. Nineteenth ESV Conference
Proceedings. Paper Number 05-0262. Washington,
Klinich KD, Reed MP, Ebert SM, Rupp JD. (2011).
Effects of Realistic Vehicle Seats, Cushion length,
and Lap Belt Geometry on Child ATD Kinematics.
Klinich KD, Reed MP, Manary MA, Orton NR,
Rupp JD (2011). Optimizing Protection for Rear
Seat Occupants: Assessing Booster Performance with
Realistic Belt Geometry Using the Hybrid III 6YO
ATD. UM-2011-40
Klinich KD, Reed MP, Ritchie N L, Manary MA,
Schneider LW, Rupp JD. Assessing Child Belt Fit,
Volume II: Effect of Restraint Configuration, Booster
Seat Designs, Seating Procedure, and Belt Fit on the
Dynamic Response of the Hybrid III 10YO ATD in
Sled Tests, UM-2008-49-2
Kortchinsky T, Meyer P, Blanot S, Orliaguet G, Puget S,
Carli P. (2008). Misuse of an adult seat belt in a
7-year-old child: a source of dramatic injuries and
a plea for booster seat use. Pediatr Emerg Care.
Kumaresan S, Yoganandan N, Pintar FA, Maiman DJ,
Kuppa S. (2000). Biomechanical study of pediatric human cervical spine: a finite element approach.
Journal of Biomechanical Engineering 122:60-71.
Kumaresan S, Yoganandan N, Pintar FA, Mueller W.
One, three and six year old pediatric cervical spine
finite element models. (1998). Frontiers in Head and
Neck Trauma. IOS Press, Amsterdam, pp 509-523.
Lalonde S, Legault F, Pedder J (2003). Relative
degradation of safety to children when automotive
restraint systems are misused. ESV Paper No. 03-0085.
Langwieder K, Hummel T, Felsch B, Klanner W. Injury
risks of children in cars— epidemiology and effect of
child restraint systems. XXIII FISITA Congress. Vol.
1. Associazione Tecnica dell’Automobile, Turin, pp
905-919, 1990.
Legault F, Gardner W, Vincent A. (1997). The effect of
top tether strap configurations on child restraint
performance. SAE 973304. Child Occupant Protection
2nd Symposium. Society of Automotive Engineers,
Warrendale, PA, pp 93-122.
January - March 2012
Lesire P, Cuny S, Alonso F, Tejera G, Cataldi M. (2007).
Misuse of child restraint systems in crash situations: dangers and possible consequences. AAAM 51:
Louman-Gardiner K, Mulpuri K, Perdio A, Tredwell S,
and Cripton, PA (2008). Pediatric lumbar chance
fractures in British Columbia: Chart review and
analysis of the use of shoulder restraints in MVAs
Accident Analysis and Prevention 40:1424–1429
Lund, AK (1986). Voluntary seat belt use among US
drivers: Geographic, socioeconomic and demographic
variation. Accident Analysis & Prevention 18(1):43-50.
Maltese MR, Locey CM, Jermakian JS, Nance ML,
Arbogast KB. (2007). Injury causation scenarios in
belt-restrained nearside child occupants. Stapp Car
Crash Journal. Vol 51:299-311.
Manary MA, Reed MP, Klinich KD, Ritchie NL, and
Schneider LW. (2006). The Effects of Tethering RearFacing Child Restraint Systems on ATD Responses.
Marriner PC, Woolford JG, Baines BA, Dance DM.
(1995). Abdominal shield booster cushions in
motor vehicle accidents. Canadian Multidisciplinary
Road Safety Conference IX. University of Montreal,
Transportation Research Center, 233-240.
Mayrose J, Priya A (2008). The safest seat: effect of
seating position on occupant mortality. J Safety Res
39(4):433-6. Epub 2008 Aug 5.
McDougall A, Brown J, Beck B, Bilston LE (2011). The
effect of varied seat belt anchorage locations on booster
seat sash guide effectiveness. ESV Paper Number
McElhaney JH, Myers BS. (1993). Biomechanical aspects
of cervical trauma. Accidental Injury: Biomechanics and
Prevention. Springer- Verlag, New York, pp 311-361.
McGrath N, Fitzpatrick P, Okafor I, Ryan S, Hensey O,
Nicholson AJ. (2010). Lap belt injuries in children. Ir
Med J 103(7):216-8.
Meissner U, Stephens G, Alfredson L. (1994). Children in
restraints. Association for the Advancement of
Automotive Medicine 38th Conference. AAAM, Des
Plaines, IL, pp 93- 106, 1994.
Melvin JW. (1995). Injury assessment reference values for
the CRABI 6-month infant dummy in a rear-facing
infant restraint with airbag deployment. SAE 950872.
Society of Automotive Engineers, Warrendale, PA.
Menon RA, Ghati YS, Roberts D. (2005). Performance
evaluation of various high back booster seats tested at
56 kph. ESV Paper No. 05-0366
Morgan C. (1999). Effectiveness of lap/shoulder belts
in the back outboard seating positions. DOT HS 808
945. National Highway Traffic Safety Administration,
Washington, DC.
Mulpuri K, Reilly CW, Perdios A, Tredwell SJ, Blair GK
(2007). The Spectrum of Abdominal Injuries
Associated with Chance Fractures in Pediatric Patients
Eur J Pediatr Surg 17:322–327DOI 10.1055/s-2007965587
Muszynski C, Yoganandan, N, Pintar, F, Gennarelli, T
(2005). Risk of Pediatric Head Injury after Motor
Vehicle Accicents. J Neurosurg. (Pediatrics 4) 102:374379.
Myers BS, Winkelstein BA. (1995). Epidemiology,
classification, mechanism, and tolerance of human
cervical spine injuries. Critical Reviews in Biomedical
Engineering 23:307-409.
Nagase H, Yonetani M, Uetani Y, Nakamura H. (2002).
Effects of child seats on the cardiorespiratory function
of newborns. Pediatr Int. 44(1):60-3.
Nance ML, Kallan MJ, Arbogast KB, Park MS, Durbin
DR, Winston, FK. (2010). Factors Associated With
Clinically Significant Head Injury in Children Involved
in Motor Vehicle Crashes. Traffic Injury Prevention
National Highway Traffic Safety Administration (2002).
Safety Rating Program for Child Restraint Systems
Docket NHTSA-2001-1053, US Department of
Transportation, Washington DC.
National Highway Traffic Safety Administration (2006).
Consumer Information Rating Program for Child
Restraints Docket NHTSA-2006-25344, US
Department of Transportation, Washington DC.
National Highway Traffic Safety Administration (2009).
Traffic Safety Facts 2008, DOT HS 811-170, National
Center for Statistics and Analysis, Washington DC.
National Highway Traffic Safety Administration (May
2010). Seat Belt Use in 2009—Use Rates in the States
and Territories Traffic Safety Facts, DOT HS 811 324
National Highway Traffic Safety Administration (2011).
Consumer Information: Program for Child Restraint
Systems, Request for Comments Federal Register
February 25, 2011, 76 FR 10637.
NHTSA (May 2009). Child restraint use in 2008-overall
results. DOT HS 811 135 Traffic Safety Facts
NHTSA Report to Congress, School Bus Safety:
Crashworthiness Research, April 2002, http://
Olson CM, Cummings, P, Rivara FP (2006). Association
of First- and Second-Generation Airbags with Front
Occupant Death in Car Crashes: A Matched Cohort
Study. American Journal of Epidemiology 164(2):161169.
O’Neil J, Daniels DM, Talty JL, Bull MJ (2009). Seat belt
misuse among children transported in belt-positioning
booster seats. Accid Anal Prev 41(3):425-9. Epub
2009 Jan 29.
Orzechowski KM, Edgerton EA, Bulas DI, McLaughlin
PM, Eichelberger MR. (2003). Patterns of injury
to restrained children in side impact motor vehicle
crashes: the side impact syndrome. J Trauma.
Park CK, Morgan RM, Digges KH. (2011). Analysis of
child dummy responses and CRS performance in
frontal NCAP tests. ESV Paper Number 11-0142
Pedder J and Hillebrandt D (2007). The development and
application of a child restraint usability rating system.
ESV Paper Number 07-0509.
Quinones-Hinojosa A, Jun P, Manley GT, Knudson MM,
Gupta N. (2005). Airbag Deployment and Improperly
Restrained Children: A Lethal Combination. Journal of
Trauma 59:729–733.
Reed M P, Ebert SM, Klinich KD, Manary MA (2008).
Assessing Child Belt Fit, Volume I: Effects of vehicle
seat and belt geometry on belt fit for children with and
without belt-positioning booster seats, UM 2008-49-1
Reed MP, Ebert SM, Sherwood CP, Klinich KD, Manary
MA. (2009). Evaluation of the static belt fit provided
by belt-positioning booster seats. Accident Analysis and
Prevention, 41:598-607.
Rouhana S. Biomechanics of abdominal trauma. (1993).
Accidental Injury: Biomechanics and Prevention.
Springer-Verlag, New York, pp 391-428.
Rutledge R, Thomason M, Oller D, Meredith W, Moylan
J, Clancy T, Cunningham P, Baker C. (1991). The
spectrum of abdominal injuries associated with the use
of seat belts. Journal of Trauma 31:820-826.
SafeKids Worldwide (2007). SafeKids Buckle Up: Ten Years
and Growing SafeKids Worldwide, Washington, DC
SafeRide News (2009). The School Bus Safety Handbook:
Choosing and Using Child Safety Restraints Systems
SafeRide News Publications, Edmonds, WA.
SafeRide News (2010). Hooked On LATCH Published
CEU Presentation, SafeRide News Publications,
Edmonds, WA.
SafeRide News (2011). The LATCH Manual 2011,
SafeRide News Publications, Edmonds, WA.
UMTRI Research Review
Seacrist T, Balasubramanian S, Garcis-Espana JF, Maltese
MR, Lopez-Valdes FJ, Kent RW, Tanji H, Higuchi
K, Arbogast KB. (2010). Kinematic Comparison
of Pediatric Human Volunteers and the Hybrid III
6-Year-Old Anthropomorphic Test Device. Annals of
Advances in Automotive Medicine. Vol 54. Oct:97-108.
Shelness A, Jewett J. (1983). Observed misuse of child
restraints. SAE 831665. Child Injury and Restraint
Conference. Society of Automotive Engineers,
Warrendale, PA, 207-215.
Sherwood C, Abdelilah Y, Crandall J. (2005). The effect
of Swedish tethers on the performance of rear-facing
child restraints in frontal crashes. ESV Paper Number
Sherwood CP, Gopalan Y, Abdelilah RJ, Crandall JR.
(2005). Vehicle Interior Interactions and Kinematics
of Rear Facing Child Restraints in Frontal Crashes.
Annual Proceedings of the Association for Advanced
Automotive Medicine. 49:215-28
Side Airbag Out-of-Position Injury Technical Working
Group (2003). Recommended procedures for evaluating occupant injury risk from deploying side airbags.
Transport Canada (2011). Booster Seat Testing.
Snyder RG, Schneider LW, Owings CL, Reynolds HM,
Golomb DH, Schork MA (1977). Anthropometry of
Infants, Children, and Youths to Age 18 for Product
Safety Design Final Report 1977 HSRI, Ann Arbor,
Society of Automotive Engineers. (1994). Securing
child restraint systems in motor vehicles. SAE J1819 (R).
Warrendale, PA, SAE.
Stalnaker RL. (1993). Spinal cord injuries to children
in real world accidents. SAE 933100. Child Occupant
Protection. Society of Automotive Engineers,
Warrendale, PA, 173-183.
Standards Australia/Standards New Zealand (2011). Child
restraint system for use in motor vehicles. AS/NZD
States JD, Huelke DF, Dance M, Green RN. (1987).
Fatal injuries caused by underarm use of shoulder belts.
Journal of Trauma 27:740-745.
Sullivan LK, Louden AE (2009). NHTSA’S initial
evaluation of child side impact test procedures. ESV
Paper No. 09-0539
Sullivan LK, Chambers FK. (1994). Evaluation of devices
to improve shoulder belt fit. DOT/HS 808 383. National
Highway Traffic Safety Administration, Vehicle
Research and Test Center, East Liberty, Ohio.
Sullivan LK, Louden AE, and Echemendia CG (2011).
NHTSA’S evaluation of a potential child side impact
test procedures. ESV Paper No.11-0227
January - March 2012
Tai A, Bilston LE, Brown J. (2011). The cumulative
effect of multiple forms of minor incorrect use in
forward facing child restraints on head injury risk.
Paper Number 11-0118.
Tsai YD, Perel M (2009). Driver’s mistakes when
installing child seats DOT HS 811 234 NHTSA,
Washington DC.
Trifiletti LB, Shields W, Bishai D, et al. (2006). Tipping
the scales: obese children and child safety seats.
Pediatrics 117(4):1197-202.
Trosseille X, Tarriere C. (1993). Neck injury criteria for
children from real crash reconstructions. SAE 933103.
Child Occupant Protection. Society of Automotive
Engineers, Warrendale, PA, pp 209-218.
Tylko S (2011). Interactions of rear facing child restraints
with the vehicle interior during frontal crash tests ESV
Paper Number 11-0406
Tylko S and Dalmotas D (2005). Protection of Rear
Seat Occupants in Frontal Crashes ESV Paper Number
Versace J. (1971) A review of the severity index. SAE
Viano DC, Parenteau CS. (2008). Fatalities of
children 0-7 years old in the second row. Traffic Inj
Prev 9(3):231-7.
Weber K, Dalmotas D, Hendrick B. (1993). Investigation
of dummy response and restraint configuration
factors associated with upper spinal cord injury in a
forward-facing child restraint. SAE 933101. Child
Occupant Protection. Society of Automotive Engineers,
Warrendale, PA, pp 185-193, 1993.
Weber K, Melvin JW. (1983). Injury potential with
misused child restraining systems. SAE 831604. 27th
Stapp Car Crash Conference. Society of Automotive
Engineers, Warrendale, PA, pp 53-59, 1983.
Weber K. (1990). Comparison of car-bed and
rearfacing infant restraint systems. 12th International
Technical Conference on Experimental Safety Vehicles.
Vol. 1. National Highway Traffic Safety Administration,
Washington, DC, pp 61-66.
Williams AF. (1976). Observed child restraint use in
automobiles. American Journal of Diseases of Children
Whitman GR, Brown KA, Cantor A, D’Aulerio LA,
Eisentraut DK, Markushewski ML. (1997). Boosterwith-shield restraint case studies. SAE 973307.
Child Occupant Protection 2nd Symposium. Society of
Automotive Engineers, Warrendale, PA, 149-157.
Wiacek C, Rudd R, Collins LA. (2011). Real world
analysis of rear seat occupant safety in frontal crashes.
ESV 11- 0380.
Zaloshnja E, Miller TR, Hendrie D. (2007). Effectiveness
of child safety seats vs safety belts for children aged 2
to 3 years. Arch Pediatr Adolesc Med 161(1):65-8.
Zonfrillo MR, Elliott MR, Flannagan CAC, Durbin DR.
(2011). Association between weight and risk of
crash-related injuries for children in child restraints.
Pediatrics 2011.
UMTRI Research Review
NNon-Profit Organization
U.S. Postage
2901 Baxter Rd
Ann Arbor MI 48109-2150
Need to change your address?
Go to
Ann Arbor, MI
Permit No. 144
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF