- Email: [email protected]
An evaluation of the effectiveness of forward facing child restraint systems
Accident Analysis and Prevention 36 (2004) 585–589
An evaluation of the effectiveness of forward facing child restraint systems Kristy B. Arbogast a,∗ , Dennis R. Durbin a,b , Rebecca A. Cornejo b , Michael J. Kallan b , Flaura K. Winston a a
The Children’s Hospital of Philadelphia, 34th and Civic Center Blvd., 3535 TraumaLink, 10th Floor, Philadelphia, PA 19104, USA b The Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia, PA, USA Received 10 February 2003; received in revised form 23 February 2003; accepted 25 February 2003
Abstract The objective of this study was to determine the effectiveness of forward facing child restraint systems (FFCRS) in preventing serious injury and hospitalization to children 12–47 months of age as compared with similar age children in seat belts. Data were obtained from a cross-sectional study of children aged 12–47 months in crashes of insured vehicles in 15 states, with data collected via insurance claims records and a telephone survey. Effectiveness estimates were limited to those children between 12 and 47 months of age seated in the back row(s) of vehicles, restrained in FFCRS, regardless of misuse, or seat belts of all types and usage. Completed survey information was obtained on 1207 children, representing 12,632 children in 11,619 crashes between 1 December 1998 and 31 May 2002. Serious injuries occurred to 0.47% of all 12–47-month olds studied, including 1.72% of those in seat belts and 0.39% of those in child restraint systems. The risk of serious injury was 78% lower for children in FFCRS than in seat belts (odds ratio (OR) = 0.22, 95% confidence interval (CI) = 0.11–0.45, P < 0.001). The risk of hospitalization was 79% lower for children in FFCRS than in seat belts (OR = 0.21, 95% CI = 0.09–0.50, P = 0.001). There was no difference between the restraint types in preventing minor injuries. As compared with seat belts, CRS are very highly effective in preventing serious injuries and hospitalization, respectively. This effectiveness estimate is substantially higher than older estimates, demonstrating the benefits of current CRS designs. These results provide those educating parents and caregivers population-based data on the importance of child restraint use. © 2003 Elsevier Ltd. All rights reserved. Keywords: Motor vehicle safety; Child safety seat; Seat belt; Effectiveness
1. Introduction Child restraint systems (CRS) have long been recommended as best practice for protecting child occupants less than 40 lb (American Academy of Pediatrics, 1999; Winston and Durbin, 1999; Weber, 2002) This recommendation has been based, in part, on an analysis by Kahane (1986) of laboratory sled tests, observational studies, and police reported crash data from the early 1980s that estimated that correctly used CRS reduce the risk of death and injury by approximately 70% as compared with children who were unrestrained. The engineering tests documented the biomechanical benefits of the CRS in spreading the crash forces over the shoulders and hips and controlling the excursion of the head and face during a crash. The study further quan-
∗ Corresponding author. Tel.: +1-215-590-3118; fax: +1-215-590-5425. E-mail address: [email protected] (K.B. Arbogast).
0001-4575/$ – see front matter © 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0001-4575(03)00065-4
tified the effectiveness of a partially misused CRS at 45% reduction in risk of fatality and serious injury. Kahane’s CRS effectiveness estimates were based on data gathered from 1974 to 1984. Further, his effectiveness estimates were based primarily on children less than 25 months of age. Since then, there have been tremendous changes in CRS and vehicle design. In addition, the concentrated focus of education on the proper use of CRS has extended the age range over which they are used and influenced how these restraints are used in actual practice. Lastly, the number of children less than 5 years of age who are unrestrained has dropped dramatically from 54% in 1984 (Kahane, 1986) to 9% in 2001 (National Highway Traffic Safety Administration, 2001). This change in distribution suggests that when evaluating the effectiveness of CRS, children of similar age in seat belts are a more contemporary comparison group. Recent attention has been placed on child occupant protection as many states have looked to upgrade their child
586
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
passenger safety laws, and the National Highway Traffic Safety Administration plans an enhancement to the Federal Motor Vehicle Safety Standard that governs child restraints. It is important to re-evaluate the previous effectiveness estimate using a data source that includes the current fleet of vehicles, CRS, and use patterns. The effectiveness estimates reported by Kahane for fatality reduction have been updated on more contemporary datasets by Partyka (1988) and Hertz (1996), but no one has evaluated the benefits of CRS in reducing serious injury since the original estimation by Kahane. Therefore, the objective of this study was to calculate the effectiveness of forward facing CRS (FFCRS) in preventing serious injury and hospitalization to children 12–47 months of age as compared with similar age children in seat belts.
2. Methods Data collected from 1 December 1998 to 31 May 2002 as part of Partners for Child Passenger Safety (PCPS) research program form the basis of this analysis. Detailed descriptions of the study population and methods involved in data collection and analysis have been previously published (Durbin et al., 2001). In summary, PCPS consists of a large-scale, population-based, child-specific crash surveillance system: insurance claims from State Farm Insurance Co. (Bloomington, IL, USA) function as the source of subjects, with telephone survey and on-site crash investigations serving as the primary sources of data. The telephone interviews provide data for a surveillance system used to describe characteristics of the population including risk factors for injury, while the crash investigations provide detailed mechanisms of injury. Vehicles qualifying for inclusion in the surveillance system were those involving at least one child occupant ≤15 years of age riding in a model year 1990 or newer State Farm-insured vehicle. Qualifying crashes were limited to those that occurred in 15 states and the District of Columbia, representing three large regions of the USA (East: NY, NJ, PA, DE, MD, VA, WV, NC, DC; Midwest: OH, MI, IN, IL; West: CA, NV, AZ). For this specific study, analyses were limited to crashes involving children aged 12–47 months riding in either seat belts or forward facing child restraints in the rear seats of vehicles in towaway crashes. Data were limited to towaway crashes so that they can be best compared with the previous effectiveness estimates calculated from the National Automotive Sampling System, a database of police reported towaway crashes across the USA. Pickup trucks were excluded from the analyses as the rear seats in these vehicle pose unique occupant safety risks. On a daily basis, data from qualifying and consenting claims were transferred electronically from all involved State Farm field offices to researchers at The Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine (CHOP/Penn). Data in this initial transfer included contact information for the insured, the ages and genders
of all child occupants, and a coded variable describing the medical treatment received by all child occupants. A stratified cluster sample was designed in order to select vehicles (the unit of sampling) for the conduct of a telephone survey with the driver. If a vehicle was sampled, the “cluster” of all child occupants in that vehicle was included in the survey. Drivers of sampled vehicles were contacted by phone and screened via an abbreviated survey to verify the presence of at least one child occupant with an injury. All vehicles with at least one child who screened positive for injury and a 10% random sample of vehicles in which all child occupants screened negative for injury were selected for a full interview. To maintain the representativeness of the full population, 2.5% of cases in which no child received care went directly to a full interview. The full interview involved a 30 min 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 6 days. 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) score. The ability of parents to accurately distinguish AIS 2 or greater injuries from those less severe using this instrument has been previously validated for all body regions of injury (Durbin et al., 1999). For the purpose of this study, serious injury was defined as any AIS 2 or greater injury excluding concussions. Concussions were excluded due to the difficulty in diagnosing this injury in the age group of children studied. AIS 2 or greater injuries include serious head and brain injuries, all internal organ injuries, spinal cord injuries, and extremity fractures. Minor injuries were defined as all abrasions, contusions, and lacerations. Restraint status of children was determined from the telephone survey. Questions related to restraint status required the respondent to describe characteristics of the restraint and how it was used. Based on the patterns of responses, we could construct the type of restraint used rather than relying on a parent’s knowledge of the product names. Among the 164 children for whom paired information on restraint use was available from both the telephone survey and crash investigations, agreement was 89% between the driver report and the crash investigator (κ = 0.74, P < 0.0001). The effectiveness estimates were limited to crashes in which the vehicle was towed from the scene of the crash, as determined from the insurance claims data. Within this group, crash severity was determined by the presence of intrusion into the occupant compartment as reported by the driver. Comparison was made between FFCRS with the full range of use characteristics and belts of all types and usage. Logistic regression modeling was used to compute the odds ratio (OR) of injury for those seated in FFCRS versus seat belts, both unadjusted and adjusted for several potential confounders including crash severity, vehicle type, and age of the child. To compute P-values and 95% confidence
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
intervals (CI) to account for the stratification of subjects by medical treatment, clustering of subjects by vehicle, and the disproportional probability of selection, Taylor Series linearization estimates of the logistic regression parameter variance were calculated using SAS-callable SUDAAN® : Software for the Statistical Analysis of Correlated Data, version 8.0 (Research Triangle Institute, Research Triangle Park, NC, USA, 2001). The Institutional Review Boards of both The Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine approved the conduct of this project.
3. Results Complete survey information was obtained on 2663 children 12–47 months of age, representing 42,476 children in 38,331 crashes. In this population, 99% were restrained with the distribution as follows: 83% FFCRS, 11% booster seats, and 6% in seat belts. In order to compare more directly with Kahane’s estimates, the effectiveness analyses were restricted to towaway crashes involving children aged 12–47 months seated in the rear row(s) of vehicles, restrained in FFCRS or seat belts. Children in pickup trucks and those
587
restrained in booster seats or who were unrestrained were excluded from this analysis. Complete survey information was obtained on 1207 children who met all inclusion criteria, representing 12,632 children in 11,619 crashes. Eleven thousand nine hundred and eighty-one (95%) of these children were restrained in FFCRS, 651 (5%) in seat belts. Derivation of this study sample is shown in Fig. 1. Table 1 provides the distribution of child age, impact direction, driver age, crash severity, and vehicle type for those restrained in seat belts and in FFCRS. Children in seat belts were more likely to be older within the age range studied (P < 0.001) than those in FFCRS. There were no differences between the two restraint types in direction of impact and crash severity proxies, such as percent of drivers under the age of 25 years and percent of crashes with intrusion. Serious injuries occurred to 0.48% (95% CI = 0.37– 0.63%) of 12–47-month olds studied, including 2.00% (95% CI = 1.06–3.74%) of those in seat belts and 0.40% (95% CI = 0.30–0.54%) of those in child restraint systems. The risk of serious injury was 80% lower for children in FFCRS than in seat belts (unadjusted OR = 0.20, 95% CI = 0.10–0.40, P < 0.001). Adjusting for crash severity, vehicle type, and age of the child yields an estimated adjusted reduction in risk of 71% (OR = 0.29, 95% CI = 0.12–0.68,
Fig. 1. Derivation of study sample. Effectiveness calculations are based on towaway crashes involving children aged 12–47 months seated in the rear row(s) of vehicles, restrained in FFCRS or seat belts—shown in the shaded boxes. Sample sizes given are based on data weighted based on probability of selection. Table 1 Child and crash characteristics for children 12–47 months of age restrained in FFCRS and in seat belts in towaway crashes
Frontal impact Side impact Rear impact Other impact 12–23 Months 24–35 Months 36–47 Months Driver <25 years of age Intrusion Passenger car SUV Minivan Large van
FFCRS, weighted percent (unweighted, n = 1091)
Seat belt, weighted percent (unweighted, n = 116)
55.6 18.8 15.7 9.8 37.9 36.0 26.1 14.9 18.8 54.5 17.8 24.9 2.9
62.3 20.4 13.6 3.7 0.9 16.8 82.2 27.5 20.6 67.1 4.5 28.4 0.0
(570) (187) (238) (96) (420) (390) (281) (174) (334) (637) (181) (255) (18)
(60) (27) (23) (6) (4) (28) (84) (28) (37) (85) (10) (21) (0)
P-value under null hypothesis of no difference between children 12–47 months of age restrained in seat belts versus FFCRS.
P-value 0.23
<0.001
0.08 0.80 0.002
588
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
P = 0.005). There was no difference between the restraint types in preventing minor injuries. 20.9% of those in seat belts and 18.1% of those in child restraint systems sustained minor injury (OR = 0.83, 95% CI = 0.43–1.60, P = 0.58). The analyses were repeated for the outcome of hospitalization. 0.31% (95% CI = 0.22–0.43%) of all 12–47-month olds studied were hospitalized, including 1.38% (95% CI = 0.62–3.03%) of those in seat belts and 0.25% (95% CI = 0.17–0.36%) of those in child restraint systems. The unadjusted risk of hospitalization was 82% lower for children in FFCRS than in seat belts (OR = 0.18, 95% CI = 0.07–0.43, P < 0.001). Adjusting for crash severity, vehicle type, and age of the child resulted in little change in this estimate of effectiveness (adjusted OR = 0.17, 95% CI = 0.05–0.51, P = 0.002). Eighty percent of the FFCRS were identified as misused, and of this group, nearly 6% had gross misuse (either the child restraint was not attached to the vehicle or the harness was not used on the child). The injury risk in those FFCRS with gross misuse (0.75%) was about twice as high as the injury risk in those FFCRS without gross misuse (0.38%), but this difference was not significant (unadjusted OR = 1.96, 95% CI = 0.64–6.04, P = 0.24). 4. Discussion This paper provides current population-based estimates of the effectiveness of FFCRS as compared with seat belts in preventing serious injuries and hospitalization to children 12–47 months of age. Our data identified that for towaway crashes, crashes of sufficient severity to render the vehicles unable to be driven from the scene of the crash, the risk of serious injury to children in FFCRS is extremely low (0.40%). Our data demonstrate that current CRS provide a 82% reduction in hospitalization, 80% reduction in serious injuries, and no reduction in minor injuries as compared with children of similar age using all types of belts. The previous effectiveness estimates reported that CRS reduced the risk of serious injury by 67% when correctly used, and 45% when partially misused as compared with unrestrained children (Kahane, 1986). Our effectiveness estimate is similar to Kahane’s previous estimate, despite the fact that our comparison group was children restrained in belts, while Kahane’s comparison group was unrestrained children. The benefits of current designs are also supported by the fact that the number of deaths to children under 5 years of age is at its lowest since record keeping began in 1975 (National Highway Traffic Safety Administration, 2002). Due to changes in CRS practice, the current study has a limited number of unbelted children, thus preventing an assessment of the effectiveness of FFCRS using unbelted children as the comparison group. The effectiveness of FFCRS in preventing serious injury as calculated in this study was evident even with a high level of child restraint misuse (80%) that is comparable to
other studies (Bull et al., 1988; Decina and Knoebel, 1997; Hummel et al., 1997). The fact that there was no difference between the restraint groups in preventing minor injury suggests that most modes of misuse lead to minor injuries, but are not critical in preventing serious injuries. One exception to this suggestion is gross misuse (either the child restraint was not attached to the vehicle or the harness was not used on the child). The percentage of seats installed with a gross misuse was very small (6%), however, those children in FFCRS with gross misuse were twice as likely to sustain a serious injury. That result was not statistically significant most likely due to sample size limitations. Several limitations in the interpretation of our results must be considered. The surveillance system is limited to children occupying model year 1990 and newer vehicles insured in 15 states and the District of Columbia. Thus, to the extent that older or uninsured vehicles differ substantially from newer insured vehicles with regard to the performance of restraints for children, results of this study may not be generalizable to occupants of these vehicles. Nearly all of the surveillance information was obtained via telephone interview with the driver/parent of the child and is potentially subject to recall bias. As noted previously, on-going comparison of parent-reported restraint use and seating position to evidence from crash investigations has demonstrated excellent accuracy of the parent report. Although one case of fatal injury was identified in our sample, injuries of this severity are likely underrepresented in our study sample. Further, the inclusion of belts of all types and use characteristics (lap only as well as lap–shoulder) as well as FFCRS with all misuse modes limits the calculation of the benefits of a properly used FFCRS when compared directly to a lap–shoulder belt. As more data become available future analyses will examine the effects of these usage modes on the effectiveness estimates. These data provide strong evidence of the effectiveness of FFCRS in preventing serious injury and provide a basis for current recommendations on the benefits of child restraint use for pre-school age children. This information can be used by public health professionals in their education tools, by legislators as evidence for improved child restraint laws, and by enforcement agencies to effect improved compliance with the laws. Those educating parents and caregivers on the importance of proper use of child restraints should encourage them to use a FFCRS for their children to its maximum weight limit, which is 40 lb for most current seats on the market, although some are going to 65 lb with use of a top tether. The transition to the next form of restraint, a belt-positioning booster seat, should not be initiated until the child has completely outgrown the FFCRS in order to maximize safety.
Acknowledgements The authors would like to thank Kathleen Weber, formerly of the University of Michigan, for her review of the
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
manuscript and State Farm Insurance Companies for their financial support of this work through the Partners for Child Passenger Safety project. In addition, the authors would like to thank the many dedicated claim representatives and personnel from State Farm, the Research Team on the Partners for Child Passenger Safety project and at TraumaLink who devoted countless hours to this study, and the parents who generously agreed to participate in the study. References American Academy of Pediatrics, 1999. Selecting and Using the Most Appropriate Car Safety Seats for Growing Children: Guidelines for Counseling Parents (RE9618). American Academy of Pediatrics. http://www.aap.org/family/01352.htm. Bull, M.J., Stroup, K.B., Gerhart, S., 1988. Misuse of car safety seats. Pediatrics 81 (1), 98–101. Decina, L.E., Knoebel, K.Y., 1997. Child safety seat misuse patterns in four states. Accid. Anal. Prev. 29 (1), 125–132. Durbin, D.R., Winston, F.K., Applegate, S.M., Moll, E.K., Holmes, J.H., 1999. Development and validation of the Injury Severity Assessment Survey/Parent Report: a new injury severity assessment survey. Arch. Pediatr. Adolesc. Med. 153, 404–408. Durbin, D.R., Bhatia, E., Holmes, J., Shaw, K., Werner, J., Sorenson, W., Winston, F., 2001. Partners for Child Passenger Safety: a unique child-
589
specific crash surveillance system. Accid. Anal. Prev. 33, 407– 412. Hertz, E., 1996. NHTSA Research Note: Revised Estimates of Child Restraint Effectiveness. National Highway Traffic Safety Administration, Washington, DC. Hummel, T., Langwieder, K., Finkbeiner, F., Hell, W., 1997. Injury risks, misuse rates and the effect of misuse depending on the kind of child restraint system. In: Proceedings of the Symposium on Child Occupant Protection, Society of Automotive Engineers, Orlando, FL. Kahane, C., 1986. An Evaluation of Child Passenger Safety: The Effectiveness and Benefits of Safety Seats. National Highway Traffic Safety Administration, Washington, DC. National Highway Traffic Safety Administration, 2001. Traffic Safety Facts 2001: Children. National Highway Traffic Safety Administration, Washington, DC. http://www-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/ TSF2001/2001children.pdf. National Highway Traffic Safety Administration, 2002. 2001 Highway Fatality Statistics. National Highway Traffic Safety Administration, Washington, DC. http://www.nhtsa.dot.gov/nhtsa/announce/press/ pressdisplay.cfm?year=2002&filename=pr55-02.html. Partyka, S., 1988. Lives Saved by Child Restraints from 1982 to 1987. National Highway Traffic Safety Administration, Washington, DC. Weber, K., 2002. Child passenger protection. In: Nahum, A., Melvin, J. (Eds.), Accidental Injury: Biomechanics and Prevention. SpringerVerlag, New York, pp. 523–549. Winston, F., Durbin, D., 1999. Buckle up! is not enough: enhancing protection of the restrained child. J. Am. Med. Assoc. 281 (22), 2070– 2072.
An evaluation of the effectiveness of forward facing child restraint systems Kristy B. Arbogast a,∗ , Dennis R. Durbin a,b , Rebecca A. Cornejo b , Michael J. Kallan b , Flaura K. Winston a a
The Children’s Hospital of Philadelphia, 34th and Civic Center Blvd., 3535 TraumaLink, 10th Floor, Philadelphia, PA 19104, USA b The Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia, PA, USA Received 10 February 2003; received in revised form 23 February 2003; accepted 25 February 2003
Abstract The objective of this study was to determine the effectiveness of forward facing child restraint systems (FFCRS) in preventing serious injury and hospitalization to children 12–47 months of age as compared with similar age children in seat belts. Data were obtained from a cross-sectional study of children aged 12–47 months in crashes of insured vehicles in 15 states, with data collected via insurance claims records and a telephone survey. Effectiveness estimates were limited to those children between 12 and 47 months of age seated in the back row(s) of vehicles, restrained in FFCRS, regardless of misuse, or seat belts of all types and usage. Completed survey information was obtained on 1207 children, representing 12,632 children in 11,619 crashes between 1 December 1998 and 31 May 2002. Serious injuries occurred to 0.47% of all 12–47-month olds studied, including 1.72% of those in seat belts and 0.39% of those in child restraint systems. The risk of serious injury was 78% lower for children in FFCRS than in seat belts (odds ratio (OR) = 0.22, 95% confidence interval (CI) = 0.11–0.45, P < 0.001). The risk of hospitalization was 79% lower for children in FFCRS than in seat belts (OR = 0.21, 95% CI = 0.09–0.50, P = 0.001). There was no difference between the restraint types in preventing minor injuries. As compared with seat belts, CRS are very highly effective in preventing serious injuries and hospitalization, respectively. This effectiveness estimate is substantially higher than older estimates, demonstrating the benefits of current CRS designs. These results provide those educating parents and caregivers population-based data on the importance of child restraint use. © 2003 Elsevier Ltd. All rights reserved. Keywords: Motor vehicle safety; Child safety seat; Seat belt; Effectiveness
1. Introduction Child restraint systems (CRS) have long been recommended as best practice for protecting child occupants less than 40 lb (American Academy of Pediatrics, 1999; Winston and Durbin, 1999; Weber, 2002) This recommendation has been based, in part, on an analysis by Kahane (1986) of laboratory sled tests, observational studies, and police reported crash data from the early 1980s that estimated that correctly used CRS reduce the risk of death and injury by approximately 70% as compared with children who were unrestrained. The engineering tests documented the biomechanical benefits of the CRS in spreading the crash forces over the shoulders and hips and controlling the excursion of the head and face during a crash. The study further quan-
∗ Corresponding author. Tel.: +1-215-590-3118; fax: +1-215-590-5425. E-mail address: [email protected] (K.B. Arbogast).
0001-4575/$ – see front matter © 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0001-4575(03)00065-4
tified the effectiveness of a partially misused CRS at 45% reduction in risk of fatality and serious injury. Kahane’s CRS effectiveness estimates were based on data gathered from 1974 to 1984. Further, his effectiveness estimates were based primarily on children less than 25 months of age. Since then, there have been tremendous changes in CRS and vehicle design. In addition, the concentrated focus of education on the proper use of CRS has extended the age range over which they are used and influenced how these restraints are used in actual practice. Lastly, the number of children less than 5 years of age who are unrestrained has dropped dramatically from 54% in 1984 (Kahane, 1986) to 9% in 2001 (National Highway Traffic Safety Administration, 2001). This change in distribution suggests that when evaluating the effectiveness of CRS, children of similar age in seat belts are a more contemporary comparison group. Recent attention has been placed on child occupant protection as many states have looked to upgrade their child
586
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
passenger safety laws, and the National Highway Traffic Safety Administration plans an enhancement to the Federal Motor Vehicle Safety Standard that governs child restraints. It is important to re-evaluate the previous effectiveness estimate using a data source that includes the current fleet of vehicles, CRS, and use patterns. The effectiveness estimates reported by Kahane for fatality reduction have been updated on more contemporary datasets by Partyka (1988) and Hertz (1996), but no one has evaluated the benefits of CRS in reducing serious injury since the original estimation by Kahane. Therefore, the objective of this study was to calculate the effectiveness of forward facing CRS (FFCRS) in preventing serious injury and hospitalization to children 12–47 months of age as compared with similar age children in seat belts.
2. Methods Data collected from 1 December 1998 to 31 May 2002 as part of Partners for Child Passenger Safety (PCPS) research program form the basis of this analysis. Detailed descriptions of the study population and methods involved in data collection and analysis have been previously published (Durbin et al., 2001). In summary, PCPS consists of a large-scale, population-based, child-specific crash surveillance system: insurance claims from State Farm Insurance Co. (Bloomington, IL, USA) function as the source of subjects, with telephone survey and on-site crash investigations serving as the primary sources of data. The telephone interviews provide data for a surveillance system used to describe characteristics of the population including risk factors for injury, while the crash investigations provide detailed mechanisms of injury. Vehicles qualifying for inclusion in the surveillance system were those involving at least one child occupant ≤15 years of age riding in a model year 1990 or newer State Farm-insured vehicle. Qualifying crashes were limited to those that occurred in 15 states and the District of Columbia, representing three large regions of the USA (East: NY, NJ, PA, DE, MD, VA, WV, NC, DC; Midwest: OH, MI, IN, IL; West: CA, NV, AZ). For this specific study, analyses were limited to crashes involving children aged 12–47 months riding in either seat belts or forward facing child restraints in the rear seats of vehicles in towaway crashes. Data were limited to towaway crashes so that they can be best compared with the previous effectiveness estimates calculated from the National Automotive Sampling System, a database of police reported towaway crashes across the USA. Pickup trucks were excluded from the analyses as the rear seats in these vehicle pose unique occupant safety risks. On a daily basis, data from qualifying and consenting claims were transferred electronically from all involved State Farm field offices to researchers at The Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine (CHOP/Penn). Data in this initial transfer included contact information for the insured, the ages and genders
of all child occupants, and a coded variable describing the medical treatment received by all child occupants. A stratified cluster sample was designed in order to select vehicles (the unit of sampling) for the conduct of a telephone survey with the driver. If a vehicle was sampled, the “cluster” of all child occupants in that vehicle was included in the survey. Drivers of sampled vehicles were contacted by phone and screened via an abbreviated survey to verify the presence of at least one child occupant with an injury. All vehicles with at least one child who screened positive for injury and a 10% random sample of vehicles in which all child occupants screened negative for injury were selected for a full interview. To maintain the representativeness of the full population, 2.5% of cases in which no child received care went directly to a full interview. The full interview involved a 30 min 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 6 days. 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) score. The ability of parents to accurately distinguish AIS 2 or greater injuries from those less severe using this instrument has been previously validated for all body regions of injury (Durbin et al., 1999). For the purpose of this study, serious injury was defined as any AIS 2 or greater injury excluding concussions. Concussions were excluded due to the difficulty in diagnosing this injury in the age group of children studied. AIS 2 or greater injuries include serious head and brain injuries, all internal organ injuries, spinal cord injuries, and extremity fractures. Minor injuries were defined as all abrasions, contusions, and lacerations. Restraint status of children was determined from the telephone survey. Questions related to restraint status required the respondent to describe characteristics of the restraint and how it was used. Based on the patterns of responses, we could construct the type of restraint used rather than relying on a parent’s knowledge of the product names. Among the 164 children for whom paired information on restraint use was available from both the telephone survey and crash investigations, agreement was 89% between the driver report and the crash investigator (κ = 0.74, P < 0.0001). The effectiveness estimates were limited to crashes in which the vehicle was towed from the scene of the crash, as determined from the insurance claims data. Within this group, crash severity was determined by the presence of intrusion into the occupant compartment as reported by the driver. Comparison was made between FFCRS with the full range of use characteristics and belts of all types and usage. Logistic regression modeling was used to compute the odds ratio (OR) of injury for those seated in FFCRS versus seat belts, both unadjusted and adjusted for several potential confounders including crash severity, vehicle type, and age of the child. To compute P-values and 95% confidence
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
intervals (CI) to account for the stratification of subjects by medical treatment, clustering of subjects by vehicle, and the disproportional probability of selection, Taylor Series linearization estimates of the logistic regression parameter variance were calculated using SAS-callable SUDAAN® : Software for the Statistical Analysis of Correlated Data, version 8.0 (Research Triangle Institute, Research Triangle Park, NC, USA, 2001). The Institutional Review Boards of both The Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine approved the conduct of this project.
3. Results Complete survey information was obtained on 2663 children 12–47 months of age, representing 42,476 children in 38,331 crashes. In this population, 99% were restrained with the distribution as follows: 83% FFCRS, 11% booster seats, and 6% in seat belts. In order to compare more directly with Kahane’s estimates, the effectiveness analyses were restricted to towaway crashes involving children aged 12–47 months seated in the rear row(s) of vehicles, restrained in FFCRS or seat belts. Children in pickup trucks and those
587
restrained in booster seats or who were unrestrained were excluded from this analysis. Complete survey information was obtained on 1207 children who met all inclusion criteria, representing 12,632 children in 11,619 crashes. Eleven thousand nine hundred and eighty-one (95%) of these children were restrained in FFCRS, 651 (5%) in seat belts. Derivation of this study sample is shown in Fig. 1. Table 1 provides the distribution of child age, impact direction, driver age, crash severity, and vehicle type for those restrained in seat belts and in FFCRS. Children in seat belts were more likely to be older within the age range studied (P < 0.001) than those in FFCRS. There were no differences between the two restraint types in direction of impact and crash severity proxies, such as percent of drivers under the age of 25 years and percent of crashes with intrusion. Serious injuries occurred to 0.48% (95% CI = 0.37– 0.63%) of 12–47-month olds studied, including 2.00% (95% CI = 1.06–3.74%) of those in seat belts and 0.40% (95% CI = 0.30–0.54%) of those in child restraint systems. The risk of serious injury was 80% lower for children in FFCRS than in seat belts (unadjusted OR = 0.20, 95% CI = 0.10–0.40, P < 0.001). Adjusting for crash severity, vehicle type, and age of the child yields an estimated adjusted reduction in risk of 71% (OR = 0.29, 95% CI = 0.12–0.68,
Fig. 1. Derivation of study sample. Effectiveness calculations are based on towaway crashes involving children aged 12–47 months seated in the rear row(s) of vehicles, restrained in FFCRS or seat belts—shown in the shaded boxes. Sample sizes given are based on data weighted based on probability of selection. Table 1 Child and crash characteristics for children 12–47 months of age restrained in FFCRS and in seat belts in towaway crashes
Frontal impact Side impact Rear impact Other impact 12–23 Months 24–35 Months 36–47 Months Driver <25 years of age Intrusion Passenger car SUV Minivan Large van
FFCRS, weighted percent (unweighted, n = 1091)
Seat belt, weighted percent (unweighted, n = 116)
55.6 18.8 15.7 9.8 37.9 36.0 26.1 14.9 18.8 54.5 17.8 24.9 2.9
62.3 20.4 13.6 3.7 0.9 16.8 82.2 27.5 20.6 67.1 4.5 28.4 0.0
(570) (187) (238) (96) (420) (390) (281) (174) (334) (637) (181) (255) (18)
(60) (27) (23) (6) (4) (28) (84) (28) (37) (85) (10) (21) (0)
P-value under null hypothesis of no difference between children 12–47 months of age restrained in seat belts versus FFCRS.
P-value 0.23
<0.001
0.08 0.80 0.002
588
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
P = 0.005). There was no difference between the restraint types in preventing minor injuries. 20.9% of those in seat belts and 18.1% of those in child restraint systems sustained minor injury (OR = 0.83, 95% CI = 0.43–1.60, P = 0.58). The analyses were repeated for the outcome of hospitalization. 0.31% (95% CI = 0.22–0.43%) of all 12–47-month olds studied were hospitalized, including 1.38% (95% CI = 0.62–3.03%) of those in seat belts and 0.25% (95% CI = 0.17–0.36%) of those in child restraint systems. The unadjusted risk of hospitalization was 82% lower for children in FFCRS than in seat belts (OR = 0.18, 95% CI = 0.07–0.43, P < 0.001). Adjusting for crash severity, vehicle type, and age of the child resulted in little change in this estimate of effectiveness (adjusted OR = 0.17, 95% CI = 0.05–0.51, P = 0.002). Eighty percent of the FFCRS were identified as misused, and of this group, nearly 6% had gross misuse (either the child restraint was not attached to the vehicle or the harness was not used on the child). The injury risk in those FFCRS with gross misuse (0.75%) was about twice as high as the injury risk in those FFCRS without gross misuse (0.38%), but this difference was not significant (unadjusted OR = 1.96, 95% CI = 0.64–6.04, P = 0.24). 4. Discussion This paper provides current population-based estimates of the effectiveness of FFCRS as compared with seat belts in preventing serious injuries and hospitalization to children 12–47 months of age. Our data identified that for towaway crashes, crashes of sufficient severity to render the vehicles unable to be driven from the scene of the crash, the risk of serious injury to children in FFCRS is extremely low (0.40%). Our data demonstrate that current CRS provide a 82% reduction in hospitalization, 80% reduction in serious injuries, and no reduction in minor injuries as compared with children of similar age using all types of belts. The previous effectiveness estimates reported that CRS reduced the risk of serious injury by 67% when correctly used, and 45% when partially misused as compared with unrestrained children (Kahane, 1986). Our effectiveness estimate is similar to Kahane’s previous estimate, despite the fact that our comparison group was children restrained in belts, while Kahane’s comparison group was unrestrained children. The benefits of current designs are also supported by the fact that the number of deaths to children under 5 years of age is at its lowest since record keeping began in 1975 (National Highway Traffic Safety Administration, 2002). Due to changes in CRS practice, the current study has a limited number of unbelted children, thus preventing an assessment of the effectiveness of FFCRS using unbelted children as the comparison group. The effectiveness of FFCRS in preventing serious injury as calculated in this study was evident even with a high level of child restraint misuse (80%) that is comparable to
other studies (Bull et al., 1988; Decina and Knoebel, 1997; Hummel et al., 1997). The fact that there was no difference between the restraint groups in preventing minor injury suggests that most modes of misuse lead to minor injuries, but are not critical in preventing serious injuries. One exception to this suggestion is gross misuse (either the child restraint was not attached to the vehicle or the harness was not used on the child). The percentage of seats installed with a gross misuse was very small (6%), however, those children in FFCRS with gross misuse were twice as likely to sustain a serious injury. That result was not statistically significant most likely due to sample size limitations. Several limitations in the interpretation of our results must be considered. The surveillance system is limited to children occupying model year 1990 and newer vehicles insured in 15 states and the District of Columbia. Thus, to the extent that older or uninsured vehicles differ substantially from newer insured vehicles with regard to the performance of restraints for children, results of this study may not be generalizable to occupants of these vehicles. Nearly all of the surveillance information was obtained via telephone interview with the driver/parent of the child and is potentially subject to recall bias. As noted previously, on-going comparison of parent-reported restraint use and seating position to evidence from crash investigations has demonstrated excellent accuracy of the parent report. Although one case of fatal injury was identified in our sample, injuries of this severity are likely underrepresented in our study sample. Further, the inclusion of belts of all types and use characteristics (lap only as well as lap–shoulder) as well as FFCRS with all misuse modes limits the calculation of the benefits of a properly used FFCRS when compared directly to a lap–shoulder belt. As more data become available future analyses will examine the effects of these usage modes on the effectiveness estimates. These data provide strong evidence of the effectiveness of FFCRS in preventing serious injury and provide a basis for current recommendations on the benefits of child restraint use for pre-school age children. This information can be used by public health professionals in their education tools, by legislators as evidence for improved child restraint laws, and by enforcement agencies to effect improved compliance with the laws. Those educating parents and caregivers on the importance of proper use of child restraints should encourage them to use a FFCRS for their children to its maximum weight limit, which is 40 lb for most current seats on the market, although some are going to 65 lb with use of a top tether. The transition to the next form of restraint, a belt-positioning booster seat, should not be initiated until the child has completely outgrown the FFCRS in order to maximize safety.
Acknowledgements The authors would like to thank Kathleen Weber, formerly of the University of Michigan, for her review of the
K.B. Arbogast et al. / Accident Analysis and Prevention 36 (2004) 585–589
manuscript and State Farm Insurance Companies for their financial support of this work through the Partners for Child Passenger Safety project. In addition, the authors would like to thank the many dedicated claim representatives and personnel from State Farm, the Research Team on the Partners for Child Passenger Safety project and at TraumaLink who devoted countless hours to this study, and the parents who generously agreed to participate in the study. References American Academy of Pediatrics, 1999. Selecting and Using the Most Appropriate Car Safety Seats for Growing Children: Guidelines for Counseling Parents (RE9618). American Academy of Pediatrics. http://www.aap.org/family/01352.htm. Bull, M.J., Stroup, K.B., Gerhart, S., 1988. Misuse of car safety seats. Pediatrics 81 (1), 98–101. Decina, L.E., Knoebel, K.Y., 1997. Child safety seat misuse patterns in four states. Accid. Anal. Prev. 29 (1), 125–132. Durbin, D.R., Winston, F.K., Applegate, S.M., Moll, E.K., Holmes, J.H., 1999. Development and validation of the Injury Severity Assessment Survey/Parent Report: a new injury severity assessment survey. Arch. Pediatr. Adolesc. Med. 153, 404–408. Durbin, D.R., Bhatia, E., Holmes, J., Shaw, K., Werner, J., Sorenson, W., Winston, F., 2001. Partners for Child Passenger Safety: a unique child-
589
specific crash surveillance system. Accid. Anal. Prev. 33, 407– 412. Hertz, E., 1996. NHTSA Research Note: Revised Estimates of Child Restraint Effectiveness. National Highway Traffic Safety Administration, Washington, DC. Hummel, T., Langwieder, K., Finkbeiner, F., Hell, W., 1997. Injury risks, misuse rates and the effect of misuse depending on the kind of child restraint system. In: Proceedings of the Symposium on Child Occupant Protection, Society of Automotive Engineers, Orlando, FL. Kahane, C., 1986. An Evaluation of Child Passenger Safety: The Effectiveness and Benefits of Safety Seats. National Highway Traffic Safety Administration, Washington, DC. National Highway Traffic Safety Administration, 2001. Traffic Safety Facts 2001: Children. National Highway Traffic Safety Administration, Washington, DC. http://www-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/ TSF2001/2001children.pdf. National Highway Traffic Safety Administration, 2002. 2001 Highway Fatality Statistics. National Highway Traffic Safety Administration, Washington, DC. http://www.nhtsa.dot.gov/nhtsa/announce/press/ pressdisplay.cfm?year=2002&filename=pr55-02.html. Partyka, S., 1988. Lives Saved by Child Restraints from 1982 to 1987. National Highway Traffic Safety Administration, Washington, DC. Weber, K., 2002. Child passenger protection. In: Nahum, A., Melvin, J. (Eds.), Accidental Injury: Biomechanics and Prevention. SpringerVerlag, New York, pp. 523–549. Winston, F., Durbin, D., 1999. Buckle up! is not enough: enhancing protection of the restrained child. J. Am. Med. Assoc. 281 (22), 2070– 2072.