Durbin DR, Elliott MR, Winston FK. Belt-Positioning Booster Seats and Reduction in Risk of Injury Among Children in Vehicle Crashes. JAMA. 2003;289(21):2835-2840. doi:10.1001/jama.289.21.2835
Author Affiliations: Department of Pediatrics, Children's Hospital of Philadelphia (Drs Durbin and Winston), and the Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine (Dr Elliott), Philadelphia.
Context Although more than a dozen states have ratified laws that require booster
seats for children older than 4 years, most states continue to have child
restraint laws that only cover children through age 4 years. Lack of booster
seat effectiveness data may be a barrier to passage of stronger child restraint
Objectives To quantify the association of belt-positioning booster seats compared
with seat belts alone and risk of injury among 4- to 7-year-old children and
to assess patterns of injury among children in booster seats vs seat belts.
Design, Setting, and Population Cross-sectional study of children aged 4 to 7 years in crashes of insured
vehicles in 15 states, with data collected via insurance claims records and
a telephone survey. A probability sample of 3616 crashes involving 4243 children,
weighted to represent 56 593 children in 48 257 crashes was collected
between December 1, 1998, and May 31, 2002.
Main Outcome Measure Parent report of clinically significant injuries.
Results Injuries occurred among 1.81% of all 4- to 7-year-olds, including 1.95%
of those in seat belts and 0.77% of those in belt-positioning booster seats.
The odds of injury, adjusting for child, driver, crash, and vehicle characteristics,
were 59% lower for children aged 4 to 7 years in belt-positioning boosters
than in seat belts (odds ratio, 0.41; 95% confidence interval, 0.20-0.86).
Children in belt-positioning booster seats had no injuries to the abdomen,
neck/spine/back, or lower extremities, while children in seat belts alone
had injuries to all body regions.
Conclusion Belt-positioning booster seats were associated with added safety benefits
compared with seat belts to children through age 7 years, including reduction
of injuries classically associated with improper seat belt fit in children.
Advocates have long recommended belt-positioning booster seats for children
who have outgrown their child safety seats.1,2 A
belt-positioning booster, either with or without a high back, raises the child
up to improve the fit of both the lap and shoulder portions of the seat belt.
Rapid, "jackknife" bending about or sliding beneath a poorly positioned vehicle
seat belt increases the risk of intra-abdominal and spinal cord injuries,
also known as "seat belt syndrome," as well as injuries to the face and brain
due to impact of the head with the child's knees or the vehicle interior.3- 8
Previously, we demonstrated the effectiveness of child restraints, including
booster seats, compared with seat belts among 2- to 5-year-old children.6 Given the relatively limited sample size of children
in booster seats at the time of that study, the effectiveness of booster seats
for 4- to 7-year-old children could not be assessed. Furthermore, belt-positioning
booster seats were combined with shield-type booster seats in analyses. These
2 seats function differently, and belt-positioning boosters are considered
by the American Academy of Pediatrics to be the optimal form of restraint
when children outgrow their child safety seats.2
Initial data on the benefits of child restraints helped to inform discussions
in many states regarding upgrades to child restraint laws to include the use
of booster seats. To date, 15 states have passed laws to include the use of
booster seats for children older than 4 years.9 Individual
states have chosen various upper age limits as a requirement for booster seat
use, ranging from 6 to 8 years. Specific data on the effectiveness of booster
seats for children older than 5 years might encourage more uniformity in these
upgraded laws. Therefore, the objective of this study was to assess the relative
effectiveness of belt-positioning booster seats compared with seat belts alone
in reducing risk of injury to children 4 to 7 years of age. In addition, we
sought to examine differences in patterns of injury among children in boosters
vs seat belts.
Data were collected from December 1, 1998, through May 31, 2002. A description
of the study methods has been published previously.10 The
project consists of a large-scale, child-specific crash surveillance system;
insurance claims from State Farm Insurance Co (Bloomington, Ill) function
as the source of subjects, with telephone survey and on-site crash investigations
serving as the primary sources of data.
Vehicles qualifying for inclusion were State Farm insured, model year
1990 or newer, and involved in a crash with at least 1 child occupant no more
than 15 years of age. Qualifying crashes were limited to those that occurred
in 15 states and the District of Columbia, representing 3 large regions of
the United States (East: New York, New Jersey, Pennsylvania, Delaware, Maryland,
Virginia, West Virginia, North Carolina, District of Columbia; Midwest: Ohio,
Michigan, Indiana, Illinois; West: California, Nevada, Arizona). After policyholders
consented to participate in the study, limited data were transferred electronically
to researchers at the Children's Hospital of Philadelphia and University of
Pennsylvania. Data in this initial transfer included contact information for
the insured, age and sex of all child occupants, and a coded variable describing
the level of medical treatment received by all child occupants as reported
by the policyholder (no treatment, physician's office or emergency department
only, hospital admission, or death).
A stratified cluster sample was designed to select vehicles (the unit
of sampling) for the conduct of a telephone survey with the driver. Vehicles
containing children who received medical treatment after the crash were oversampled
so that the majority of injured children would be selected while maintaining
the representativeness of the overall population. If a vehicle was sampled,
all child occupants in that vehicle were included in the survey. Drivers of
sampled vehicles were contacted by telephone and, if a passenger had received
medical treatment, screened via an abbreviated survey to verify the presence
of at least 1 child occupant with an injury. All vehicles with at least 1
child who had a positive screen for injury and a 10% random sample of vehicles
in which all child occupants who were reported to receive medical treatment
but screened negative for injury were selected for a full interview; a 2.5%
sample of crashes in which no medical treatment was received were also selected.
The full interview involved a 30-minute telephone survey with the driver of
the vehicle and parent(s) of the involved children. Only adult drivers and
parents were interviewed. The median length of time between the date of the
crash and the completion of the interview was 10 days, with 95% of interviews
completed within 51 days of the crash.
The eligible study population consisted of all 367 020 children
riding in 246 245 State Farm–insured vehicles newer than 1990 reporting
a crash claim between December 1, 1998, and May 31, 2002. Claim representatives
correctly identified 95% of eligible vehicles, and 74% of policyholders consented
to participation in this study. Of these, 19% were sampled for interview and
an estimated 81% of these were successfully interviewed (Figure 1). Comparing the included sample with known population values
from State Farm claims, we observed little difference: in both the sample
and the population, 40%, 36%, and 24% of the vehicles were located in the
East, Midwest, and West regions, respectively, and 58% of the sampled vehicles
were model 1996 or newer. In the sample, 53% were passenger cars, 20% minivans,
18% sport utility vehicles, 6% pickup trucks, and 2% large passenger/cargo
vans compared with 55%, 18%, 18%, 7%, and 2% in the population, respectively;
32% were nondriveable after the crash compared with 30% of the population.
The mean age of the child occupants in the sample was 6.9 years compared with
7.3 years in the population.
For cases in which child occupants were seriously injured or killed,
in-depth crash investigations were performed. Cases were screened via telephone
to confirm the details of the crash. Contact information from selected cases
was then forwarded to a crash investigation firm (Dynamic Science Inc, Annapolis,
Md), and a full-scale on-site crash investigation was conducted using custom
child-specific data collection forms. Among cases selected for investigation,
97% were completed. For the purposes of this analysis, these cases were used
to examine the validity of information obtained from the telephone survey.
Restraint status of children was determined from the telephone survey.
Children were classified as unrestrained or restrained, with the restraint
type further classified as seat belt, belt-positioning booster, shield booster,
or child safety seat. Among the 161 children for whom paired information on
restraint use was available from both the telephone survey and crash investigations,
agreement was 88% between the driver report and the crash investigator (κ
= 0.74; P<.001). Crash severity was categorized
both by the towaway status of the vehicle (ie, whether the vehicle was towed
from the crash scene), as indicated in the insurance claims data, and by driver
report via the telephone survey of intrusion into the occupant compartment
of the vehicle. Seating location of each child was determined from the telephone
survey. Among the 170 children for whom paired information on seating position
(front vs rear) was available from both the telephone survey and crash investigations,
agreement was 99% between the driver report and the crash investigator (κ
= 0.99; P<.001).
Survey questions regarding injuries to children were designed to provide
responses that were classified by body region and severity based on the Abbreviated
Injury Scale (AIS) score11 and have been previously
validated for their ability to distinguish AIS scores of 2 or more from less
severe injuries.12 For the purposes of this
study, children were classified as injured if a parent/driver reported a clinically
significant injury, ie, any injury with an AIS score of 2 or greater (concussions
and more serious brain injuries, all internal organ injuries, spinal cord
injuries, and extremity fractures) or facial lacerations.
Separate oral consent was obtained from eligible participants for the
transfer of claim information from State Farm to The Children's Hospital of
Philadelphia and University of Pennsylvania for the conduct of the telephone
survey and the crash investigation. The study protocol was reviewed and approved
by the institutional review boards of both The Children's Hospital of Philadelphia
and the University of Pennsylvania School of Medicine.
The primary purpose of these analyses was to compute the relative risk
of injury for children aged 4 to 7 years restrained in belt-positioning booster
seats compared with seat belts. Age 4 years was chosen as the lower bound
for our analyses because current recommendations for optimal restraint indicate
that children younger than 4 years should be restrained in child safety seats.1,2 χ2 Tests of association
were used to compute P values based on the null hypothesis
of no association between restraint type and risk of injury. Logistic regression
modeling was used to compute odds ratios (ORs) of injury for those seated
in belt-positioning booster seats vs seat belts, both unadjusted and adjusted
for several potential confounders, including differences in driver age (<25
vs ≥25 years), seating position (front vs rear), crash severity, and vehicle
Because sampling was based on the likelihood of an injury, subjects
least likely to be injured were underrepresented in the study sample in a
manner potentially associated with the predictors of interest.13 To
account for this potential bias and to adjust inference to account for the
stratification of subjects by medical treatment and clustering of subjects
by vehicle, robust χ2 tests of association and Taylor series
linearization estimates of the logistic regression parameter variances were
calculated using SAS-callable SUDAAN version 7.5 (Research Triangle Institute,
Research Triangle Park, NC). Results of logistic regression modeling are expressed
as unadjusted and adjusted ORs with corresponding 95% confidence intervals
(CIs). Because of the small number of injuries to certain body regions, differences
in risk of injury by body region were assessed using a nonparametric permutation
test14 based on the null hypothesis of no injury
risk difference by restraint type by body region.
This analysis is restricted to the 3616 crashes involving 4243 children
aged 4 to 7 years, weighted to represent 56 593 children in 48 257
crashes. Among all 4- to 7-year-olds, 3519 were using either a seat belt or
belt-positioning booster seat, representing 45 701 children, or 81% of
all 4- to 7-year-olds in the study population. As expected, restraint use
of the children varied by age. Seat belts were used by 42% of 4-years-olds,
72% of 5-year-olds, and 89% of 6- and 7-year-olds; belt-positioning booster
seats were used by 16% of 4-year-olds, 13% of 5-year-olds, and 4% of 6- and
7-year-olds. The majority of children in belt-positioning booster seats (81%)
were in high-backed booster seats.
Table 1 provides the distribution
of driver age, seating row, crash severity, and vehicle type among 4- to 7-year-olds
restrained in seat belts (unweighted n = 3282; weighted n = 40 389) and
in belt-positioning booster seats (unweighted n = 237; weighted n = 5312).
Children in belt-positioning booster seats were less likely to be in vehicles
driven by younger drivers and more likely to be rear-seated than children
in seat belts.
Injuries occurred in 1.81% of all 4- to 7-year-olds, including 1.95%
of those in seat belts and 0.77% of those in belt-positioning booster seats.
There were 5 deaths identified among children in seat belts, while no children
in belt-positioning booster seats died. The unadjusted odds of injury were
61% lower for children aged 4 to 7 years in belt-positioning boosters than
in seat belts (OR, 0.39; 95% CI, 0.20-0.77). The unadjusted odds of injury
reduction ranged from 56% among 4-year-olds (OR, 0.44; 95% CI, 0.15-1.31)
to 81% for 6-year-olds (OR, 0.19; 95% CI, 0.04-0.83). Despite the appearance
of variable effectiveness by age, there was no significant difference in the
effectiveness of booster seats by individual year of age (P = .66).
Children in seat belts were more likely to be in vehicles driven by
young drivers and more likely to be located in the front seat (Table 1), both factors that increase risk of injury. Adjustment
for these factors, together with crash severity, vehicle type, and age of
the child, yielded an estimated adjusted reduction in risk of 59% for children
in boosters (OR, 0.41; 95% CI, 0.20-0.86; P = .02).
The adjusted odds of injury reduction did not vary significantly by age (P = .44), ranging from 46% among 4-year-olds (OR, 0.54;
95% CI, 0.17-1.74) to 85% among 6-year-olds (OR, 0.15; 95% CI, 0.03-0.80).
Restricting our analysis to rear-seated children only did not substantially
alter the booster seat protective effect (adjusted OR, 0.40; 95% CI, 0.18-0.89; P = .03). Comparing booster seat users with lap-shoulder
belt users only had no effect (adjusted OR, 0.39; 95% CI, 0.17-0.86; P = .02). Airbag exposure was rare for both belted and
booster children, and adjusting for it as well had no effect (adjusted OR,
0.40; 95% CI, 0.19-0.83; P = .01).
Table 2 provides the distribution
of the body regions of injury for children in belt-positioning booster seats
and seat belts. Children in seat belts had injuries to every body region.
However, children in belt-positioning booster seats had injuries only to the
head, face, chest, and upper extremities. Of particular note among children
in belt-positioning booster seats is the lack of injuries to the abdomen and
neck/back/spine—part of the body regions characterized by the "seat
belt syndrome" constellation of injuries—although these differences
were statistically significant only for abdominal injuries.
This study confirms that belt-positioning booster seats are associated
with a reduced risk of injury compared with seat belts in children aged 4
to 7 years. As noted previously, belt-positioning booster seats function by
raising a child up on the vehicle seat so that his seated height is more like
that of an adult, allowing both portions of the belt to fit more properly.
Belt-positioning booster seats have small handles, guides, or a slot that
help to position the lap portion of the belt low and flat across a child's
upper thighs.15 Many boosters have high backs
that not only provide the child with head support but also have upper belt
guides to optimize the position of the shoulder portion of the belt.15 The bottom cushions of belt-positioning boosters
are also shallower than the vehicle seat, allowing the child's knees to bend
comfortably at the edge of the booster. This encourages a child to sit up
straight in the seat with his back against the seat back.16
Our results provide the first real-world evidence that this optimal
positioning of the belt is associated with significantly fewer injuries classically
associated with seat belt use in young children. Belt-positioning booster
seats appeared to significantly reduce the risk of injuries to the abdomen.
In addition, children in belt-positioning booster seats in our sample had
no injuries to the abdomen, spine, or lower extremities, while children in
seat belts had injuries to every body region.
At present, 15 states have passed legislation upgrading child restraint
laws to include the appropriate restraint of children aged 4 years or older
in booster seats, with several other states actively considering similar upgrades
to their child restraint laws.9 Federal legislation
has also recently been passed to encourage states to upgrade their child restraint
laws to include the use of booster seats for children aged 4 years or older.17 Parents rely on state laws for guidance about properly
restraining their children.18 Laws that are
in closer alignment with current best-practice recommendations may help reduce
confusion among parents regarding the most effective way to protect their
children. We have previously reported significant increases in booster seat
use, particularly among 4-year-old children.19 Previous
research suggests that parental knowledge of booster seats has increased during
the period of this study.20 These results may
assist state and federal policymakers by demonstrating the current level of
interest on the part of parents to optimally restrain their older children.
Our data suggest that the safety benefits of booster seats are present through
age 7 years.
We found that children in belt-positioning booster seats were more likely
to sit in the rear seats of vehicles than children in seat belts. This may
indicate that many parents link the concepts of child restraint use and seating
position or it may be an indicator of safer drivers. Rear-seated children
in booster seats derive safety benefits not only from appropriate restraint
but also by optimal seating position.21,22 Our
estimate of the effectiveness of booster seats remained unchanged when restricted
to children in the rear seats of vehicles.
Booster effectiveness was also unchanged when restricted to children
using both the lap and shoulder portion of the seat belt (as opposed to the
lap belt only). It is possible that children using booster seats as well as
those in seat belts used only the lap portion of the seat belt, either because
that is all that was available or because they placed the shoulder belt behind
their back or under their arm. We included these children in the primary analysis
to evaluate effectiveness of boosters compared with belts in the "as used"
real-world situation. We combined high-back and backless belt-positioning
booster seats for our analyses. Further research will be required to evaluate
whether one design or the other may offer an important incremental safety
This study relied on parent report of restraint use by children, which,
if differentially reported by restraint type, might have biased the results.
To determine some potential effects of restraint misclassification, we assumed
that those identified as unrestrained were indeed unrestrained and that the
observed risk of injury among the unrestrained correctly estimates their risk
for each year of age. If we further assume that all children reported as restrained
in belt-positioning booster seats were correctly classified, 50% of those
reported as restrained in seat belts would have to be unrestrained for the
risk differences between those reported as being in belt-positioning booster
seats and those reported as being in seat belts to be eliminated. Since the
maximum misclassification rate among those in seat belts (which requires the
risk of injury among those correctly reported to be 0) is 62%, this is highly
unlikely. Assuming that 20% of those reported to be in booster seats were
actually unrestrained (the maximum possible is 25%), a misclassification rate
of 60% would be required among those reported to be in seat belts to nullify
risk differences. It is highly unlikely that differential misclassifications
this large exist. Similar results were obtained in analyses stratified by
year of age.
The National Highway Traffic Safety Administration currently recommends
that all children who have outgrown child safety seats should be restrained
in booster seats until they are at least 8 years old, unless they are 57 in
(145 cm) tall.1 In the overall Partners for
Child Passenger Safety population, booster seat use by 8-year-olds is negligible;
therefore, they were not included in the analysis. The difference in height
between 50th-percentile 7- and 8-year-olds is approximately 2 in (5 cm), with
more than 95% of 8-year-olds expected to be less than 57 in (145 cm).23 Based on available data regarding the fit of seat
belts in children,16 there is no reason to
suspect that the benefits of booster seats demonstrated through age 7 years
would not remain through age 8 years. Recent data suggest that restrained
children older than age 7 years are at a higher risk of injury than younger
children in crashes, suggesting that seat belts alone may not provide optimal
protection to older children as well.24 Further
research is needed to identify characteristics of optimal restraint devices
for children of all ages.
Belt-positioning booster seats are associated with added safety benefits
over seat belts for children through age 7 years. Pediatricians should educate
parents regarding current recommendations for optimal restraint, including
the use of belt-positioning booster seats within their practice. In addition,
state child restraint laws should be revised to include the use of booster
seats for children through age 7 years. Pediatricians can play an important
role in advocating for this legislation in their state.