Ebel BE, Koepsell TD, Bennett EE, Rivara FP. Use of Child Booster Seats in Motor Vehicles Following a Community CampaignA Controlled Trial. JAMA. 2003;289(7):879-884. doi:10.1001/jama.289.7.879
Author Affiliations: Departments of Pediatrics (Drs Ebel and Rivara) and Epidemiology (Drs Koepsell and Rivara), and Harborview Injury Prevention and Research Center (Drs Ebel, Koepsell, and Rivara), University of Washington, Seattle; and Children's Hospital and Regional Medical Center, Seattle, Wash (Ms Bennett).
Context Once children have outgrown car seats, booster seats protect from injury
better than lap and shoulder belts alone. However, the majority of children
aged 4 to 8 years use only an adult seat belt.
Objective To evaluate the effectiveness of a multifaceted community booster seat
campaign in increasing observed booster seat use among child passengers in
Design Prospective, nonrandomized, controlled community intervention trial.
Setting and Participants The campaign was initiated in 4 communities in the greater Seattle,
Wash, area between January 2000 and March 2001. Eight communities in Portland,
Ore, and Spokane, Wash, served as control sites. We observed 3609 booster-eligible
children (those aged 4-8 years and weighing 18-36 kg [40-80 lb]).
Main Outcome Measure Observed booster seat use 15 months after the start of the campaign.
Results Before the campaign began, 13.3% of eligible children in the intervention
communities and 17.3% in the control communities were using booster seats,
adjusting for child age, driver seat belt use, and sex of driver. Fifteen
months after the start of the campaign, adjusted booster seat use had increased
to 26.1% in the intervention communities and 20.2% in the control communities
(P = .008 for the difference in time trends between
intervention and control communities).
Conclusion These data suggest that a multifaceted community education campaign
can significantly increase the use of child booster seats.
Motor vehicle collisions remain the leading cause of death for children
aged 4 to 8 years.1 Booster seats are more
effective than seat belts at reducing the risk of injury for these older children.2 Nonetheless, only an estimated 6% to 19% of 4- to
8-year-old children currently ride in booster seats, while the majority use
only an adult seat belt.3- 5 To
date, there are no published trials of community interventions to increase
booster seat use. In 1999, we began a 2-year community campaign to increase
booster seat use in King County, Washington. This report describes the effectiveness
of the campaign in increasing observed booster seat use.
We conducted a prospective, nonrandomized, controlled trial to evaluate
the community intervention. Booster seat use in each community was observed
at baseline just before beginning the intervention, and then observations
were repeated 15 months later. Four communities in the greater Seattle, Wash,
area served as intervention sites, comprising about 240 000 people. Eight
communities in Portland, Ore, and Spokane, Wash, served as control sites.
These communities were approximately matched on household per capita income
and population size (Table 1),6,7 and were chosen because they represented
distinct neighborhoods with clear geographic boundaries.
The campaign targeted both parents and children. It sought to increase
parental awareness of the need for booster seats, reduce the motivational
and financial barriers to purchasing a seat, and reinforce booster seat use
through public health messages delivered from many different sources. Campaign
messages were developed based on the Precede-Proceed model of behavioral change,
identifying predisposing, enabling, and reinforcing factors affecting behavior.8 Parents perceived that seat belts provided adequate
protection, increasing awareness with the campaign slogan, "Is Your Child
Ready for Seat Belt? Think Again!" To address parental confusion and lack
of knowledge, all secondary messages provided specific guidelines for booster
seat use. Enabling factors included a focus on health care providers as disseminators
of information and the use of discount coupons. A new booster seat law served
as an important reinforcing factor.
Social marketing provided the structure for campaign development.9 Audience focus was narrowed to particularly target
parents with younger children aged 3 to 5 years. The booster seat message
was clear and specific. A variety of tactics were used, including media, publicity,
education outreach, policy change, and provider education, all of which used
a consistent message and approach. Messages emphasized the benefits of booster
seat use and the consequences of not using one. Social learning theory constructs
were considered in specific program elements and approach.10 Contents
of materials were designed to build parent confidence in choosing and using
a booster seat, and to provide the knowledge parents needed to make informed
Before the campaign launch, we conducted 3 focus groups with parents
to investigate reasons for booster seat use and nonuse and test the efficacy
of our public health messages.11 Participants
were recruited from child care centers in Seattle area neighborhoods. They
all had at least 1 child aged 3 to 5 years, had mixed incomes, and may or
may not have been using a car seat or booster seat. Participants responded
to a range of questions about their booster seat knowledge, attitudes, beliefs,
and behaviors. A professional focus group facilitator conducted each session
and summarized key themes and messages. Information from the focus groups
guided message development and priorities.
We developed the King County Booster Seat Coalition to bring together
parents, public health professionals, community outreach workers, childcare
providers, law enforcement officers, physicians, emergency medical technicians,
and educators. Coalition members helped set priorities for the campaign and
spearheaded outreach efforts with child care providers, schools, and parents.
They developed classes and health fair events for families and disseminated
booster seats, materials, and discount coupons.
Based on the findings of our focus groups and in consultation with the
coalition and a parent feedback committee, we designed a multifaceted public
(Box). All educational materials provided information on
why seat belts are not adequate and how a booster seat can help position a
child in the car and make him/her more comfortable. Materials included images
of different types of booster seats, because many parents indicated that they
were confused by the varying styles. Brochures provided parenting suggestions
to address parental perceptions of peer pressure by other children to not
use a booster seat. We created public service announcements for television,
radio, and print media, targeting community newspapers and parent magazines.
An easy-to-read pamphlet was designed and translated into 7 different languages.
Specific intervention materials and programs were designed for health professionals,
child care facilities, schools, and community fairs. A booster seat telephone
hotline was set up through the Washington Safety Restraint Coalition. Callers
could request materials and coupons and could ask car-specific questions from
car-safety experts. We established a Web site (http://www.boosterseat.org),
on which parents could find information about booster seats, a fit test, request
materials or a coupon, answer commonly asked questions, find locations for
car-seat checks, and find links to the growing body of helpful Web sites with
booster seat information. The Web site has received steadily increasing traffic
with 81 000 hits per month by August 2002.12
With support from a booster seat manufacturer and a retail outlet, we
negotiated a coupon that provided $10 off the price of a low-back booster
seat, lowering the price of this seat to less than $20. These coupons were
available through child care centers, physicians' offices, community organizations,
a toll-free telephone hotline, or via our booster seat campaign Web site.
The study was approved by the institutional review boards at the University
of Washington Human Subjects Division.
Community coalition of agencies and organizations to promote the use
of booster seats
Citizen advisory group of parents and caregivers to provide feedback
on campaign messages and materials and to develop strategies to ensure community
Broad-based community education program to increase knowledge and awareness
of the importance of booster seat use, which included:
Organization and group newsletter articles
Booster seat Web site
Tip sheet, brochures, and flyers in multiple languages
Telephone information line where parents can call for materials and
with questions about booster seats and car seats
Resource kits for preschools and health care providers
Radio public service announcements
Television public service announcements
Local news reports
Educational programs to address barriers to booster seat use, including
defining types of booster seats, identifying where devices are available,
and providing alternatives for automobiles with lap-only belts
Discount booster seat coupons
Car seat training programs and in-services for health care providers,
child care providers and educators, law enforcement, emergency medical service
personnel, and child passenger safety advocates
We measured baseline booster seat use in each of the 12 communities
beginning in January 2000. Follow-up observations were conducted 15 months
later beginning in March 2001. We conducted observations at 83 child care
centers and after-school programs, as has been previously described.4 Because the booster seat promotion campaign was neighborhood-based,
these sites were chosen to measure booster seat use in the local 4- to 8-year-old
population, rather than the wider population of children one might expect
to find at a toy store or larger shopping center. Fast-food sites were pilot-tested
but were not used as observation sites because very few booster eligible-children
per hour were observed, making observations too inefficient, and because several
communities had only 1 or 2 fast-food restaurants within their borders. Random
intersection sites were not used because the study required that vehicles
be stopped and approached, so that drivers could be directly questioned about
child age and weight.
We obtained consent from the center directors and obtained oral consent
from all participating drivers. Centers were asked to withhold information
to the parents about the scheduling of the survey, so as not to alter routine
behavior. A pilot study was conducted to determine the feasibility of the
observational study and brief survey. Observers were recruited from local
colleges and communities and underwent a standard training program to explain
the aims of the study and the methodology. Practice observations were conducted
under supervision before data collection. Teams of 2 or more observers visited
each site. Cars were approached in the parking lot at pickup times only after
the driver had an opportunity to secure any children in the vehicle, and to
fasten his/her own seat belt. Cars leaving with only a nonambulatory child
were excluded from the survey. Drivers reported child's age, weight and height,
and distance from home. Trained observers looked into the car to assess actual
restraint use by the driver and all occupants.
Advocates recommend booster seats for children who have outgrown child
harness seats (usually at 18 kg [40 lb] or around age 4 years) until they
fit properly in the vehicle seat belt.13,14 Children
do not fit into the adult seat belt well until they are 143 cm tall, generally
at around 36 kg (80 lb), and at least 8 years.15 The
American Academy of Pediatrics states that a child should stay in a booster
seat until the belt fits, usually when the child reaches about 143 cm in height
and is between 8 to 12 years of age.13 Because
the majority of drivers in our survey were unable to report a child height,
raising concerns that those who did may have reported inaccurate heights,
we chose to use only age and weight criteria to define recommended booster
seat use. We defined a child who should use a booster seat as any child between
18 and 36 kg (40-80 lb) and aged between 4 to 8 years, or a child aged between
4 to 8 years when weight data were not available. Three-year-old children
weighing 18 kg (40 lb) were also characterized as being booster seat eligible,
because the majority of child harness seats have a maximum weight limit of
18 kg (40 lb). Ten children were 2 years of age and were reported to weigh
18 kg (40 lb) or more. These children were considered to be too young for
booster seat use and were classified as child harness seat users. Children
who were more than 36 kg (80 lb) or older than 8 years were considered to
be big enough for the adult lap and shoulder belt. There were no 3-year-old
children who weighed more than 36 kg (80 lb), and no 8-year-old children weighing
less than 18 kg (40 lb).
The overall effectiveness of the intervention was estimated using a
generalized linear mixed model, adjusted for child level variables (age, sex),
car level variables (driver sex, driver seat belt use), and incorporated the
effects of clustering at the levels of car, site, and community.16,17 The
adjusted prevalence of booster seat use in each community was obtained using
model-based direct adjustment,18 using the
combined population of all studied children as the standard population. Data
were analyzed with STATA version 7.0 (Stata Corp, College Station, Tex) and P≤.05 was considered statistically significant.
We observed 5656 children traveling in cars during baseline and follow-up
observations, 3609 of whom were eligible for booster seat use based on age
and weight criteria. Eighty-five percent of drivers approached agreed to our
survey; there was no significant difference in response rate by community.
Among children eligible for a booster seat, mean age was 5 years and mean
weight was approximately 23 kg (50 lb). Half of all drivers were surveyed
within 11 minutes by car from home. Eighty-three percent of drivers were wearing
a seat belt themselves. Nearly 70% of drivers were female (Table 2).
The overall effectiveness of the intervention was estimated using a
generalized linear mixed model, allowing for adjustment by child level variables
(child age) and car level variables (driver sex and seat belt use), while
incorporating the effects of nested clustering by car, site, and community.16,17Table 3 shows the adjusted prevalence of booster seat use by study
group and observation time. Adjustment was performed using model-based direct
adjustment,18 using the combined population
of all studied children as the standard population. Therefore, each adjusted
prevalence may be interpreted as the prevalence of booster seat use that would
have been observed in a certain study group at a certain time if such children
had had the same distribution of covariates (age, driver sex, driver seat
belt use) as did the study population as a whole. The adjusted prevalence
of booster seat use in intervention communities was 13.3% at baseline and
rose to 26.1% at follow-up. In the control communities, the adjusted prevalence
of baseline booster seat use was 17.3% and increased to 20.2% at follow-up
(P = .008 for the difference in time trends between
intervention and control communities), albeit with considerable variation
among neighborhoods. Child sex and distance from home were not significant
predictors or confounders of booster seat use in the final multivariate model.
Booster seat use decreased with increasing child age. At follow-up,
booster seat use was greatest for 4- to 6-year-old children (26%) and was
uncommon for children aged 7 to 8 years (9%). For each additional year of
age, the odds ratio (OR) of using a booster seat was 0.71 (95% confidence
interval [CI], 0.65-0.76), indicating that an older child was significantly
less likely to be properly restrained in a booster seat. Driver use of seat
belts was also strongly associated with correct booster seat use, with an
OR of 5.51 (95% CI, 2.86-10.65). Female drivers were more likely to be driving
child passengers correctly restrained in booster seats than were male drivers
(OR, 1.56; 95% CI, 1.04-2.33).
To our knowledge, this is the first controlled evaluation of a community
intervention designed to increase booster seat use. We found that our campaign
significantly increased booster seat use among booster-eligible children in
intervention communities relative to control communities.
Most children who are aged between 4 to 8 years, weigh between 18 and
36 kg (40-80 lb), or with height less than 143 cm do not fit well into an
adult seat belt.14,15,19 Recent
anthropometric data suggest that a child who is at least 143 cm tall and weighs
at least 36 kg (80 lb) has a median age of 11 years20;
therefore, the recommendation that booster seats be used by children who are
younger than 8 years recognizes that most 8-year-old children will meet the
height and weight criterion for booster seat use. Data from mechanical sled
tests,15,19 computer crash simulations,
and the limited epidemiological data available2,21 all
suggest that booster seats are protective for children who do not yet fit
into seat belts. Therefore, safety experts recommend that children who have
outgrown child harness seats and do not yet fit adult seat belts should use
a booster seat, generally children younger than 8 years and 36 kg (80 lb).13,14 However, this study shows that most
booster-eligible children are not using booster seats and may be inadequately
protected in a crash.
Focus group studies we conducted before beginning the booster seat campaign
found that many parents did not know about booster seats or had an inaccurate
understanding of when it is safe to move a child to a lap shoulder belt.11,22 Parents reported that they had not
received consistent information about when a child should be in a booster
seat. Other barriers to booster seat use included child resistance, peer pressure,
and difficulty finding affordable booster seats.
To encourage booster seat use, we designed a coordinated, community
education campaign using multiple avenues for getting parents information
about booster seats. Health promotion theory suggests that multicomponent
interventions are likely to be more effective at changing behavior.8,23 A similar community-based campaign
was successful at promoting the use of bicycle helmets and reducing bicycle-related
head injury.11,24,25 The
development of a coalition of community members, educators, public health
officials, and health care and child care providers continues to provide input
and guidance to the intervention. There has been continued demand for informational
brochures, as well as steadily increasing local and national traffic on the
booster seat Web site and telephone hotline.
The increasing national emphasis on booster seat use by the National
Highway Traffic Safety Administration, Safe Kids Coalitions, and others contributed
additional positive messages to parents, although presumably the control communities
in our study were exposed to these same messages. In addition, the legislatures
of Washington and Oregon both passed booster legislation during the time course
of the study, although in both instances the legislation did not go into effect
until after the intervention period, and the proposed legislation would have
affected both control and intervention sites.
Our study design allowed control for the effects of other state and
national booster seat initiatives to see the effect of a more focused community-based
campaign. Demonstrated effectiveness at increasing booster seat use may also
provide impetus for other states to invest in community campaigns and to consider
legislative changes in support of booster seat use.
Our results also demonstrated variability in neighborhood booster seat
usage patterns. We cannot say exactly why this variation occurred, but there
are multiple exogenous factors that contribute to the prevalence of prevention
behaviors in the community. We were interested in knowing the effects of the
special campaign over and above changes that might have occurred to other
information in the community, region, or nation. We are not aware of any particular
reason for the variations in booster seat prevalence. In 1 comparison community,
the prevalence of booster seat use increased quite dramatically from 21% to
54%. This was an affluent community, with high prevalence of booster seat
use during baseline observations. Data were analyzed, including this site
as a control site, and therefore the increase in booster seat use in this
community was fully accounted for in the data analysis.
This study had a number of limitations. It was conducted in 3 northwest
cities and therefore may not be generalizable to other parts of the country.
It was not a randomized study, so there is always the possibility of unmeasured
confounders. We attempted to reduce the effects of confounding by comparing
booster use in each community to baseline booster use in the same community.
The design of this study did not permit us disaggregate intervention components
to determine which were most effective. Finally, our study relied on a brief
and noninvasive survey. We were therefore not able to physically measure children
but relied instead on parental report of age, weight, and height.
In conclusion, this multifaceted community campaign was able to significantly
increase booster seat use among booster-eligible children. It shared elements
with other successful community-based injury strategies in that it had a strong
theoretical basis, involved coalition-building, relied on a focused public
health message, used focus groups to guide campaign messages and development,
and included an evaluation of campaign effectiveness. This program might serve
as a useful model for other communities working on injury-prevention strategies.