RE S E ARCH AR T I C L E The Lack of Motor Vehicle Occupant Restraint Use in Children Arriving at School KATHRYN D. EMERY, MDa S. GLENN FARIES, MDb ABSTRACT BACKGROUND: Motor vehicle crashes are the leading cause of mortality for children aged 4-14 in the United States. Many children are driven daily to school, increasing their exposure to potential injury, especially if they are not appropriately restrained. Observing the level of motor vehicle occupant restraint (MVOR) use upon school arrival could lead to valuable information for the development of an injury prevention program targeted to this population. This study compares the rate of MVOR use upon arrival to elementary schools with that at regional intersections. METHODS: One hundred thirty-five census tracts in the Denver, Colorado, area were identified and combined into 5 regions. Within each region, 1 controlled intersection and 1 elementary school were observed. Observations were conducted for 1 hour at each of these locations. RESULTS: At intersections, 618 children in private vehicles were observed. MVOR use rate was 71.2% (440/618). At schools, 665 children in private vehicles were observed. MVOR use was 24.8% (165/665). Children arriving at school were less likely to be restrained than those observed at intersections (relative risk: 0.35, 95% CI: 0.30-0.40). CONCLUSIONS: Children at elementary schools showed substantially lower MVOR use rates than those observed at controlled intersections. Further investigation is imperative to determine the reason behind the low use of MVOR in children being transported to school. Based on the reasons elicited, interventions emphasizing the importance of always using MVOR can be developed and tailored to meet the educational needs of parents transporting their children to school in private vehicles. Keywords: motor vehicle occupant safety; children; school. Citation: Emery KD, Faries SG. The lack of motor vehicle occupant restraint use in children arriving at school. J Sch Health. 2008; 78: 274-279. aAssociate Professor, (emery.kathryn@tchden.org), Department of Emergency Medicine, University of Colorado at Denver and Health Sciences Center, The Children's Hospital, Box B-251, 13123 East 16th Avenue, Aurora, CO 80045. bAssociate Professor, (faries.glenn@tchden.org), Department of Emergency Medicine, University of Colorado at Denver and Health Sciences Center, The Children's Hospital, Box B-251, 13123 East 16th Avenue, Aurora, CO 80045. Address correspondence to: Kathryn Emery, Associate Professor, (emery.kathryn@tchden.org), Department of Emergency Medicine, University of Colorado at Denver and Health Sciences Center, The Children's Hospital, Box B-251, 13123 East 16th Avenue, Aurora, CO 80045. 274 d Journal of School Health d May 2008, Vol. 78, No. 5 d 2008, American School Health Association Motor vehicle crashes (MVCs) are the leading cause of mortality for children aged 4-14 in the United States.1 The potential for injury is higherwhen restraints are used improperly or not used at all. Also, the potential exposure to MVC-related injury is greater with an increased number of motor vehicle miles traveled.2 Eighteen million children in the United States, approximately 40% of all primary and secondary school students, travel to and from school in a motor vehicle that is not a school bus. Ninety-eight percent of school-aged MVC deaths during normal school hours occur in passenger vehicles or to pedestrians, bicyclists, or motorcyclists. Yearly, 80% of the 152,000 nonfatal injuries during school hours occur in passenger vehicles.3 Unrestrained children are more likely to be injured, suffer more severe injuries, and die in MVCs than restrained children.1 Yet, studies have shown a lack of proper restraint use in children of all ages. Younger children are appropriately restrained more frequently than older children; as few as 10% of booster seat-aged children are reported in the literature as being appropriately restrained. In contrast, up to99%of infants are appropriately restrained.4-7 It has been shown that school-based interventions can be effective in changing behaviors.8-10 In a systematic review of educational programs to increase child safety seat use, the Task Force on Community Preventive Services found that ''education programs, when used alone have insufficient evidence to determine their effectiveness.'' However, the Task Force also concluded that incentive and education programs are effective in increasing child safety seat use in the short term. The programs included in this reviewwere implemented in both day care centers and community wide among a variety of target populations. An educational and enforcement campaign in Durham,North Carolina, improved correct use ofmotor vehicle occupant restraint (MVOR) at 2 elementary schools, from 36% to 64% at 1 elementary school and from 49% to 71% at a second elementary school.11 The National Occupant Protection Use Survey (NOPUS) contains information on a state-by-state basis about occupant restraints, but it is difficult to break out schools from other points of observation. Of note, guidelines for observational studies recommend schools as locations for observation.12 Our intent was to observe children in vehicles arriving at schools to help determine future interventions and sites for intervention within our metro area. We hypothesized thatMVOR use for children upon school arrival would be representative of MVORuse for children in general. Elementary schools present a unique environment for observing MVOR use and also represent a concentrated target population in which an educational intervention could be made available to parents and children. METHODS The study design is a matched pairs cross-sectional study of restraint use observed at 5 schools and 5 intersections. Subjects The subjects of the study were elementary school- aged children observed at 5 intersections and 5 elementary schools. The elementary schools were all part of the Denver Public School (DPS) system. Each elementary school and the corresponding controlled intersection were within each of 5 regions defined as described below. Permission to perform the unobtrusive observations was secured from either the principal of the school in question or the superintendent of schools for DPS. DPS consists of 151 schools in total, of which 73 are elementary schools. Procedure In order to define neighborhoods and optimize demographic similarities between neighborhoods, each of 135 census tracts was compared with the others by developing a matching ratio. Demographic data were gathered for each of the city's 184 census tracts based on the 2000 census.13 Demographic data included 18 characteristics (Table 1). For the 135 census tracts with populations of 100 or more, a matching ratio was calculated using the 18 demographic characteristics. The 18 demographic variables were normalized with respect to the median values of all census tracts. The matching ratio was calculated as the sum of the squares of the differences of the normalized measures for the 18 demographic characteristics. By comparing the matching ratios between neighborhoods, the 135 census tracts were combined into 5 regions with optimal similarity within neighborhoods. Observational surveys were conducted in accordance with National Highway Traffic Safety Administration's (NHTSA) uniform criteria for state observational surveys of seat belt use. Observational requirements specified that observers follow a predetermined clear policy, including direct observation on roadways within the state, along with seat belt use determined by use or nonuse of a shoulder belt. Instructions to observers specified which road and which direction of traffic were to be observed. Observers followed clear instructions on how to start and stop an observation period as well as how to start and stop observations if traffic flow was too heavy to observe all vehicles or if vehicles began moving too quickly for observation. For this study, training included an education session (which was delivered via videotape) on how to perform the observations and was followed by a posttest to ensure that the Journal of School Health d May 2008, Vol. 78, No. 5 d 2008, American School Health Association d 275 individual trained was capable of assessing seat belt use or nonuse in photographs specifically obtained for the purpose of the training video. This tool was developed as part of a previous NHTSA-funded study and was submitted to NHTSA with the data from that study. This tool had been validated prior to use on a separate population of observers. All members of the observational team were trained using this standardized protocol. Observations included restraint use or nonuse and vehicle type. Rear facing MVORs were excluded from observations as these would contain infants, a population that this study was not designed to include. Identification of restraint misuse was not attempted since we could not determine the child's age, weight, or height. In this study, all vehicles were able to be counted. We collected information on vehicle type and restraint use of all school-aged children within the vehicle both in the front and back seats. One controlled intersection, an intersection with a stop sign or stoplight, within each region was identified, and the same trained volunteer conducted observations for 1 hour at each of these intersections. Here, vehicles either slowed or came to a complete stop. Observations at intersections occurred during March and April 2003 for 1-hour periods during the evening rush hour and on weekend mornings. These 2 times were chosen to maximize the volume of traffic that passed through the intersections at which observations were taking place. Pilot observations at different times revealed that not enough cars with child passengers passed by to supply sufficient data. Observation sites were randomly assigned to the selected day of week and time of day periods. All vehicles were surveyed. Within each of the 5 regions, 1 elementary school was identified. School authorities were contacted requesting permission for a researcher to observe motor vehicle restraint use on one morning. Observations occurred during April and May 2003. Observations were performed during the morning drop-off time at a controlled intersection 1-2 blocks from the drop-off point of each school, inclusive of most traffic to that school. All vehicles were surveyed. The original and 2 additional trained volunteers performed these observations. Enrollment in the school lunch program was used as a proxy for socioeconomic status.14 The study was approved by the Colorado Multiple Institutional Review Board (CoMIRB). CoMIRB is housed at the University of Colorado at Denver and Health Sciences Center (UCDHSC) and has been established to review biomedical and behavioral research involving human subjects conducted at or supported by UCDHSC and several other affiliated institutions. Data Analysis Data were entered into Microsoft Excel (Microsoft, Redmond, WA). Relative risks (RR) were calculated to compare MVOR use at intersections versus MVOR use in children arriving at school. Data were analyzed in 2005. RESULTS A total of 1283 observations of MVOR use were performed. At intersections, 618 children in private vehicles were observed. Overall MVOR use rate was 71.2% (440/618), ranging from 61.5% to 100%. At the 5 schools, 665 children in private vehicles were observed. Overall MVOR rate was 24.8% (165/665), ranging from 13% to 46% (Figure 1). Children arriving at school were less likely to be restrained (RR: Table 1. Demographic Characteristics of the 5 Regions Area 1 2 3 4 5 Population 139,385 89,414 95,846 81,285 87,847 Number of HH 54,176 47,258 39,776 38,212 46,049 Number of workers 68,099 49,067 53,940 47,205 54,794 Number of people older than 25 years 86,953 60,459 62,619 56,746 63,327 Median age 33 35 35 37 38 Median HH income 28,825 24,614 45,476 49,552 46,325 Persons per HH 2.57 1.89 2.41 2.13 1.91 Percent HH with children younger than 18 years 35.1 18.9 32.8 24.3 19.1 Percent HH with married couples 39.5 19.4 39.6 43.0 36.0 Percent blue collar 48.3 43.2 39.4 34.3 31.4 Percent college degrees 19.2 32.4 35.9 42.3 46.4 Percent non-Hispanic white 43.0 47.1 48.2 81.0 80.8 Percent Hispanic 50.0 32.7 12.3 13.8 9.2 Percent black 3.4 18.8 38.2 2.1 6.7 Percent Asian, Native American 5.9 4.6 4.2 3.5 4.0 Percent HH with vehicles 83.5 70.8 89.0 92.1 93.5 Percent HH with multiple vehicles 44.9 25.7 49.0 55.1 48.4 Vehicles per capita 0.57 0.55 0.65 0.79 0.83 HH, household. 276 d Journal of School Health d May 2008, Vol. 78, No. 5 d 2008, American School Health Association 0.35, 95% CI: 0.30-0.40). Region 4 (the area with the highest MVOR use rate at both intersections and schools) had the highest median household income and the highest percentage of households with married couples. We also found that children in vans were most likely to be restrained at both schools and intersections, while children in pickup trucks were least likely to be restrained (Figure 2). By univariate analysis, there was a significant difference in restraint use at schools compared with intersections by vehicle type for automobiles, sport utility vehicles (SUV) and vans (passenger/minivans). There was no significant difference in restraint use at schools compared with intersections for pickup trucks. Pickup trucks accounted for 10% of vehicles observed. There were significantly greater proportions of SUVs and pickup trucks at schools compared with intersections. There were significantly less automobiles at schools compared with intersections. There was no significant difference in the proportion of vans at schools compared with intersections. In general, schools with a lower socioeconomic level (based on the percentage of students at that school who participate in the free lunch program) were found to have a lower rate of MVOR use compared to schools with higher socioeconomic levels (Figure 3). DISCUSSION It is imperative that major efforts are undertaken to decrease the high rates of injury and death due to Figure 1. MVOR Use at Schools and Intersections 0 10 20 30 40 50 60 70 80 90 100 Percentage of children restrained 1 2 3 4 5 overall Regions Schools Intersections Figure 2. MVOR Use Rate at Schools and Intersections by Vehicle Type 0 10 20 30 40 50 60 70 80 90 100 Percentage of children restrained SUV Auto Van Pickup Vehicle type Schools Intersections Journal of School Health d May 2008, Vol. 78, No. 5 d 2008, American School Health Association d 277 motor vehicle collisions. Many children are driven almost daily to and from school; this population is a prime target for intervention. The original aim of this study was to document MVOR use upon school arrival. Through comparison of children in vehicles bound for school with those in the same geographical area but not bound for school, we determined that school-bound children are a special population with extremely low MVOR use upon school arrival. Our data for MVOR use rate at intersections (71.2%) are consistent with individual state data from NHTSA. This organization reports a safety belt use rate of 72% for the state of Colorado.15 Children that we observed being transported to school in private vehicles were much less likely to be restrained (24.8%). Breaking the data down by vehicle type, we found that with the exception of pickup trucks, children in all other types of vehicles were more likely to be restrained at intersections than upon school arrival. The rate of restraint use in pickup trucks at both schools and intersections was extremely low, an indication that this population could be a specific target for intervention. There are some general explanations for nonuse of MVOR, which could apply both to children observed at intersections and upon school arrival. NHTSA, in their motor vehicle occupant safety survey, identified reasons for nonuse of car seats for children younger than 9 years. These reasons included the following: only short time in car, child does not like it, seat is not available, we are in a hurry, child will not stay in it.12 Adults who allow children to travel unrestrained may have differences in perception of risk, or be less riskaverse, thus being less vigilant about MVOR use. In a recent study, focus groups of drivers with booster seat-aged children revealed that there are differences in the degree to which parents are motivated by their children's safety. Differences in risk perception and parenting style were found between those who restrained their children in booster seats and thosewho restrained booster seat-aged children in adult seat belts.16 The existence of a child passenger restraint law has been shown to be a strong incentive for MVOR use.17 The law in Colorado requires that infants be transported in a rear facing seat until they are at least 1 year of age and at least 20 pounds. Children aged 1-4 and between 20 and 40 pounds must ride in an appropriate car seat. Colorado law also requires that children who are older than 4 years but younger than 6 years ride in a booster car seat unless they are more than 55 in. tall. Children younger than 16 years must be buckled in a seat belt. The infant seat, child safety seat, and seat belt provisions of the law are primary enforcement, whereas the booster seat provision is secondary enforcement. The seat belt law for adults in Colorado provides only for secondary enforcement. Unawareness of the law could be a contributing factor to nonuse of MVOR, but it would seem unlikely that this lack of awareness would explain the discrepancy between the observed rates of MVORuse at intersections versus upon school arrival in the same geographic and demographic area. It may be postulated therefore that there are issues at play, which are specific to transportation of children to and from school. Observers noted that many of the children who were in the cars were wearing backpacks, making it impossible for them to be appropriately restrained. Due to the chaos of getting children ready for school, especially when running late, the parent may be less vigilant about appropriately restraining his children. Some of the NHTSA issues short time in car, in a hurry, seat availabilitymay play a larger role in transportation of elementary school students. It is notable that in schools with a lower socioeconomic climate, children were even less likely to be restrained. The NOPUS has demonstrated lower use of restraints in rural areas and low-income communities.18 Figure 3. MVOR Use Upon School Arrival (%) Compared With the Percentage of Students in the Free Lunch Program at Each Observed School 0 20 40 60 80 100 School 1 School 2 School 3 School 4 School 5 Percentage of children restrained Percentage of children in the free lunch program 278 d Journal of School Health d May 2008, Vol. 78, No. 5 d 2008, American School Health Association Edgerton found an impressively low seat belt use rate in school-aged children of a low-income Hispanic community compared to the national average.19 There are limitations to this study. There were significant differences in the proportion of specific vehicle types at schools versus intersections. Observational surveys are prone to unmeasured confounders. Vehicles were not stopped by observers, so some children may have been missed due to tinted windows or height of vehicle. Children may have released their MVOR early in order to prepare to exit the vehicle upon school arrival. Observations were not conducted at successively farther intersections to determine when children began releasing their MVOR. Riding unrestrained is the greatest risk factor for death and injury among child occupants of motor vehicles, so observations in this study concerned MVOR use only and did not evaluate appropriate restraint use. Ages were not recorded or estimated. There could have been a greater number of preschoolers observed at intersections, potentially increasing the MVOR use rate since it has been shown that younger children are more likely to be appropriately restrained.4-7 CONCLUSIONS Children at local elementary schools showed a substantially lower MVOR use rate than those observed at controlled intersections. Further investigation is imperative to determine the reasons for this in order to develop a targeted intervention. MVOR use varies based on the environment of observation, school versus general intersection. REFERENCES 1. National Highway Traffic Safety Administration, National Center for Statistics & Analysis, Research & Development. Traffic Safety Facts 2000. Children. Washington, DC: National Highway Traffic Safety Administration; 2000. DOT HS 809 324. 2. Kweon YJ, Kockelman KM. Overall injury risk to different drivers: combining exposure, frequency, and severity models. Accid Anal Prev. 2003;35(4):441-450. 3. Transportation Research Board of the National Academies. The Relative Risk of School Travel. A National Perspective and Guidance for Community Risk Assessment. Washington, DC: Transportation Research Board; 2002. Special Report 269. Available at: http:// trb.org/publications/sr/sr269.pdf. Accessed January 2005. 4. Glassbrenner D. The Use of Child Restraint in 2002. Washington, DC: US Department of Transportation, National Highway Traffic Safety Administration; 2003. DOT HS 809 555. Available at: http://www.-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/RNotes/2003/ ChildRestRN.pdf. Accessed January 2005. 5. National Highway Traffic Safety Administration. National Survey