Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
April 2005

Injury Risk to Restrained Children Exposed to Deployed First- and Second-Generation Air Bags in Frontal Crashes

Author Affiliations

Author Affiliations: Division of Emergency Medicine (Drs Arbogast and Durbin), Division of General Pediatrics, Department of Pediatrics (Dr Winston), and The Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia (Drs Durbin and Elliott and Mr Kallan); and TraumaLink: The Interdisciplinary Pediatric Injury Control Research Center, The Children's Hospital of Philadelphia (Drs Arbogast, Durbin, and Winston).


Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005

Arch Pediatr Adolesc Med. 2005;159(4):342-346. doi:10.1001/archpedi.159.4.342

Objective  To estimate the risk of serious nonfatal injuries in frontal crashes among belted children seated in the right front seat of vehicles in which second-generation passenger air bags deployed compared with that of belted children seated in the right front seat of vehicles in which first-generation passenger air bags deployed.

Design and Setting  We enrolled a probability sample of 1781 seat belt–restrained occupants aged 3 through 15 years seated in the right front seat, exposed to deployed passenger air bags in frontal crashes involving insured vehicles in 3 large US regions, between December 1, 1998, and November 30, 2002. A telephone interview was conducted with the driver of the vehicle using a previously validated instrument. The study sample was weighted according to each subject’s probability of selection, with analyses conducted on the weighted sample.

Main Outcome Measure  Risk of serious injury (Abbreviated Injury Scale score of ≥2 injuries and facial lacerations).

Results  The risk of serious injury for restrained children in the right front seat exposed to deployed second-generation passenger air bags was 9.9%, compared with 14.9% for similar children exposed to deployed first-generation passenger air bags (adjusted odds ratio, 0.59; 95% confidence interval, 0.36-0.97).

Conclusion  This study provides evidence based on field data that the risk of injury to children exposed to deploying second-generation passenger air bags is reduced compared with earlier designs.

Beginning in the 1990s, a portion of the deaths and injuries to children in motor vehicle crashes were attributed to exposure to deploying passenger air bags. During this time several reports of children killed by deploying passenger air bags in minor crashes focused attention on the unique needs of children in automotive safety.15 Braver et al1 estimated that passenger air bags reduced adult occupant fatalities by 18% in frontal crashes and by 11% in all crashes while increasing the risk of death for children younger than 10 years by 34%. Others supported this finding.57 Passenger air bags have also been shown to increase the risk of nonfatal injury to children by 87%.8

Faced with the challenge of balancing benefits to adult passengers while minimizing risk to child occupants, the National Highway Traffic Safety Administration (NHTSA) initiated a 2-pronged response of education and regulation. First, the NHTSA, joined by national organizations such as the National Safety Council, Athe Air Bag and Seat Belt Safety Campaign, the American Academy of Pediatrics, the American College of Emergency Physicians, and the American Medical Association, recommended that all child passengers younger than 13 years sit in the rear seats of vehicles. Second, in 1997, the NHTSA enacted a substantial regulatory change to Federal Motor Vehicle Safety Standard 208, which provided automakers a choice between certifying frontal crash performance for unbelted adults by either rigid barrier tests or sled tests.9 This change in the standard, in many cases, resulted in the redesign of frontal air bags to reduce the force with which they deploy.10 These new air bags are generally referred to as “second-generation air bags.”

Child exposure to passenger air bags remains a persistent public health problem. Despite the rear seat recommendation, many children—1 of every 8 children in crashes8—continue to sit in the right front seat of passenger air bag–equipped vehicles. As the number of passenger air bag–equipped vehicles on the roadways continues to rise, the actual number of children exposed to a deploying passenger air bag concurrently increases. In early 1999, 73 of every 10 000 children in crashes were exposed to an air bag deployment. By the end of 2002, this figure rose to 148 per 10 000 children in crashes.11

To date several studies have examined the influence of the new air bag technology on adult drivers and passengers. Studies using the National Automotive Sampling System, a probability sample of police-reported crashes in the United States,12,13 and a trauma center–based surveillance system14 suggested protection to adults in high-speed crashes. As of July 2004, the NHTSA reported that second-generation air bags had resulted in substantially fewer child fatalities than earlier designs,15 but, to our knowledge, no studies have measured the effect of air bag redesign on a large sample of child passengers, nor have any studies examined the effect of second-generation air bags on the risk of nonfatal injuries. Therefore, the objective of this study was to estimate the risk of serious nonfatal injuries in frontal crashes among belted children in the front seat of vehicles in which second-generation passenger air bags deployed, compared with that of belted children in the front seat of vehicles in which first-generation passenger air bags deployed.


Data were collected from December 1, 1998, through November 30, 2002. A description of the study methods has been published previously.16 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 interview 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 being 15 years or younger. 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, Deleware, Maryland, Virginia, West Virginia, North Carolina, and the District of Columbia; Midwest: Ohio, Michigan, Indiana, and Illinois; and West: California, Nevada, and 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, the age and gender of each child occupant, and a coded variable describing the level of medical treatment received by each child occupant as reported by the policyholder (no treatment, physician's office or emergency department only, admitted to the hospital, or death).

A stratified cluster sample was designed to select vehicles (the unit of sampling) to conduct a telephone interview with the driver. Vehicles containing children who received medical treatment following the crash were oversampled so that most 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 telephone interview to verify the presence of at least 1 child occupant in the vehicle with an injury. All vehicles with at least 1 child who screened positive 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 where no medical treatment was received was also selected. The full interview with the driver of the vehicle and parent(s) of the involved children lasted approximately 30 minutes. The median length of time between the date of the crash and the completion of the interview was 6 days, with 95% of the interviews completed within 47 days of the crash.

The eligible study population consisted of all 430 308 children riding in 288 187 State Farm–insured vehicles newer than 1990 reporting a crash claim between December 1, 1998, and November 30, 2002. Claim representatives correctly identified 95% of the eligible vehicles identified by the electronic claims system, and 73% of the policyholders consented for participation in this study. Of these, 18% were sampled for interview and an estimated 81% of these were successfully interviewed (Figure). Comparing the sample with population values from all eligible State Farm claims, we see little difference in geographic region, vehicle model year, vehicle type, tow status of the vehicle, and child age.

Image not available

Derivation of overall project study sample from initially eligible population. Analyses were conducted on data from full interviews of crashes with children aged 3 through 15 years restrained and seated in the right front seat in frontal crashes where a passenger air bag deployed.

For crashes in which child occupants were seriously injured or killed, in-depth crash investigations were performed. Parents and caregivers were contacted via telephone to confirm the details of the crash. Contact information from selected crashes 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 crashes selected for investigation, 97% were completed. For this analysis, these crashes were used to examine the validity of information obtained from the telephone interview.


Restraint status of children was determined from the telephone survey. For this analysis, only those children who were restrained by a seat belt were included. 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 statistic = 0.74). Crash severity was categorized by driver report of intrusion into the occupant compartment of the vehicle via the telephone survey. 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 statistic = 0.99).

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,17 and have been previously validated for their ability to distinguish AIS 2 or greater from less severe injuries.18 For this study, children were classified as “seriously injured” if a parent or driver reported a clinically significant injury: 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. Injuries included all those injuries sustained by the child occupant, not just those injuries that could be attributed to the air bag deployment. Direction of the first impact was determined from the driver report of the plane of the vehicle where the initial impact occurred (front, side, rear, other). Only frontal crashes were included in this analysis.

Deployment of the air bag was determined from survey responses. Air bags were categorized as first-generation (all premodel year 1998 vehicles and 1998 vehicles without redesigned air bags) and second-generation air bag systems (all 1998 vehicles with redesigned systems and all 1999 and newer vehicles). Information on the date a particular vehicle was certified to the new standards was determined from data obtained from NHTSA’s Special Crash Investigation Division, which obtained the information directly from the manufacturers as part of their evaluation of the air bag regulation. 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.


Analyses assessing risk of injury were limited to frontal impact crashes with a child occupant in the right front seat who was exposed to a deployed passenger air bag. Because very few children in child restraints or booster seats or younger than 3 years were exposed to a passenger air bag, we restricted the study sample to seat belt–restrained children between 3 and 15 years. The risk of serious nonfatal injuries in frontal crashes among belted children seated in the right front seat of vehicles in which second-generation passenger air bags deployed was compared with that of belted children seated in the right front seat of vehicles in which first-generation passenger air bags deployed.

Logistic regression modeling was used to compute the odds ratio (OR) of injury for those exposed to deployed first-generation air bags vs second-generation air bags, both unadjusted and adjusted for several potential confounders including child age modeled as a 3-level covariate (3-8 years, 9-12 years, 13-15 years linear term), crash severity (intrusion or no intrusion), and vehicle type (passenger car, sport utility vehicle [SUV]), minivan, pickup truck, or cargo van).

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.19 To account for this potential bias, sampling weights were used equal to the inverse of the probability of selection. To adjust inference to account for these sampling weights, as well as 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: Software for the Statistical Analysis of Correlated Data, Version 8.0 (Research Triangle Institute, Research Triangle Park, NC). Results of logistic regression modeling were expressed as unadjusted and adjusted ORs with corresponding 95% confidence intervals (CIs).


Complete interview data were obtained on– 1148 (unweighted n) seat belt–restrained, right front-seated children ages 3 through 15 years exposed to a deploying passenger air bag in a frontal impact crash, representing 1781 (weighted n) child passengers in the study population. One thousand thirty-one children were exposed to a first-generation passenger air bag, and 750 were exposed to a second-generation passenger air bag. Characteristics of the study sample are listed in Table 1. Serious injuries (AIS score ≥2 plus facial lacerations) were reported in 210 sampled children representing an estimated 228 children or 12.8% of the population (14.9% in the first-generation group vs 9.9% in the second-generation group). After adjusting for child age, vehicle type, and crash severity, children exposed to second-generation systems sustained fewer serious injuries than those exposed to first-generation passenger air bags (adjusted OR, 0.59; 95% CI, 0.36-0.97) (Table 2). There were fewer injuries to all body regions except the abdomen in the second-generation group, although none of the body region–specific results achieved statistical significance. There were no fatalities in either group.

Table 1. 
Image not available
Child, Driver, Vehicle, and Crash Characteristics Among Restrained Children Exposed to Deploying First- and Second-Generation Passenger Air Bags in Frontal Crashes
Table 2. 
Image not available
Risk of Serious Injury in Frontal Crashes by Body Region for Restrained 3- Through 15-Year-Olds Exposed to a Deployed Second-Generation Passenger Air Bag Compared With That of Those Exposed to a Deployed First-Generation Passenger Air Bag

This study provides the first field assessment of the effect of the regulatory changes to air bags implemented in the late 1990s on restrained child passengers. Data suggest that the objective of this change in policy, the reduction of injuries to child occupants, is being met by these new systems in frontal crashes with air bag deployment.

Our previous work reported that children exposed to a deploying passenger air bag were at an 87% increased risk of serious injury compared with children in similar crashes who were not exposed to a passenger air bag.8 That study, however, combined vehicles of model year 1990 to 2002, irrespective of the changes in regulation that occurred with model year 1998. The current analyses extend our prior findings and, to our knowledge, provide the first field assessment of the recent substantial air bag regulatory changes.

We found that restrained children ages 3 through 15 seated in the right front seat exposed to a deployed second-generation air bag experience a 41% reduction in the odds of serious injury, after adjusting for differences in child age, vehicle type, and crash severity between the 2 air bag types. These data also highlighted nonsignificant trends of reduction in injuries to the head, face, and lower extremities. This is consistent with previous laboratory data on second-generation air bag designs, which predicted a reduction compared with older designs in head injury risk for occupants closest to the air bag during deployment.20 Poor seat belt and vehicle seat fit in young children often lead to close occupant proximity to the air bag during deployment.

Despite the risk reduction to restrained children exposed to a deployed second-generation air bag in this study, clinicians, educators, and policymakers should maintain the current recommendation to place children younger than 13 years in the rear rows in an age-appropriate restraint. First-generation air bags will remain in the vehicle fleet for many years and the development of educational messages for consumers that distinguish between the 2 designs is complex. Clinicians can continue to play an important role through anticipatory guidance regarding appropriate seating position for children in second-generation air bag–equipped vehicles. Future research should be conducted that explicitly compares the risk of injury to children seated in the right front seat in vehicles equipped with newer passenger air bags to those seated in the rear seat before changes to current seating recommendations are considered.

Air bag designs continue to undergo significant redesigns in an effort to optimize their effectiveness in serious crashes while minimizing their risk of adverse injuries in more minor crashes. In 2001, additional revisions were made to the federal regulation covering air bag performance now requiring the testing of air bag systems for all sized occupants, including children. The field data presented in this article provide a suitable baseline to which future air bag designs can be compared.

Our study sample represents the entire spectrum of crashes reported to an insurance company including property damage only, as well as bodily injury crashes. While our sample included a significant number of vehicles with intrusion into the occupant compartment, we likely do not have a representative sample of the most severe crashes. The NHTSA’s Special Crash Investigation Division monitors fatal child injuries associated with air bag deployments and have documented a reduction in fatalities with second-generation air bags.15 This result is likely attributed to a combination of technology change and increased rear seating. Future work should monitor the safety provided to child occupants by newer-generation air bags in the most severe crashes.

Data for this study were obtained via telephone interview primarily with the driver and/or parent of the child and were, therefore, subject to potential misclassification. Ongoing comparison of driver-reported child restraint use and seating position to evidence from crash investigations has demonstrated a high degree of agreement. In addition, our results on age-specific restraint use and seating position are similar to those of other recently reported population-based studies of child occupants.4,21 It is possible that some of those reported to be restrained by a seat belt were actually unrestrained, thus resulting in a form of selection bias in that unrestrained children would have been incorrectly included in the study sample. There is no evidence that this misreporting would vary by generation of air bag; therefore, it is difficult to estimate the potential effect of this misclassification on our results. Further, our estimate of crash severity, driver report of intrusion, is a gross proxy of that measure. More direct measures such as change in velocity were unable to be determined via telephone survey and, therefore, are unable to be used in these analyses. The effect of this potential source of measurement error, a common limitation of many crash studies, may affect our results in an unpredictable manner. It is unlikely that errors in reporting would vary between drivers of vehicles with second-generation air bags and those with first-generation air bags, thus the effect of any errors on the results of this study of seat belt restrained children only would be limited.

Lastly, these analyses were conducted on a belted population in crashes where the air bag deployed and cannot define the net hazard to children in vehicles equipped with second-generation air bags. Studies incorporating unrestrained child occupants, nondeployment crashes and nonfrontal crashes would be necessary to conclusively assess the performance on the new designs. Future work using the Partners for Child Passenger Safety database will be used to address this question.

Surveillance data of the nature presented in this study are crucial for identifying the magnitude of the risk of injury to children from seated in front of deploying passenger air bags. While the pattern of injuries noted in this study among children exposed to passenger air bags is consistent with injuries caused by interaction with air bags sustained by children who contact an air bag during a crash, this study, based on survey data, cannot not establish conclusively this causation for the subjects’ injuries. Improvements seen in the vehicles with the newer air bag designs may possibly be due to other crash-worthiness improvements in the vehicle rather than the design changes in the air bag alone. Future work will use crash investigation methods to ascertain how often the air bag in the source of the injury and to elucidate the specific mechanisms by which children are injured in these crashes.

To our knowledge, this study provides the first evidence that second-generation air bag designs provide reduced injury risk over the first-generation systems for restrained child passengers in the front seat exposed to a deployed air bag. These results offer assurance to policymakers and restraint designers regarding the effect of their changes in the late 1990s. Owing to a diversity of crash conditions and air bag designs in the current fleet, physicians offering anticipatory guidance should continue with current recommendations for children to sit age-appropriately restrained in the rear seat. Consistent maintenance of this important message is critical as consumers may misinterpret the improvements in air bag design as permission to place children in the front seat.

Back to top
Article Information

Correspondence: Kristy B. Arbogast, PhD, The Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, 3535 TraumaLink, 10th Floor, Philadelphia, PA 19104 (

Accepted for Publication: November 23, 2004.

Acknowledgment: We thank State Farm for their financial support of this work through the Partners for Child Passenger Safety project. In addition, we also thank the many dedicated claim representatives and personnel from State Farm, the Research Team on the Partners project and at TraumaLink, who devoted countless hours to this study, and the parents who generously agreed to participate in the study.

Braver  ERFerguson  SAGreene  MALund  AK Reductions in deaths in frontal crashes among right front passengers in vehicles equipped with passenger air bags. JAMA 1997;2781437- 1439
Cases from the Special Crash Investigation Program Version, National Highway Traffic Safety Administration. Available at: July 7, 2004
Winston  FReed  R Airbags and children: results of a National Highway Traffic Administration special investigation into actual crashes.  Paper presented at: 40th Stapp Car Crash Conference November 4, 1996 Albuquerque, NM
Wittenberg  ENelson  TGraham  J The effect of passenger airbags on child seating behavior in motor vehicles. Pediatrics 1999;1041247- 1250
Graham  JGoldie  SSegui-Gomez  M  et al.  Reducing risks to children in vehicles with passenger air bags. Pediatrics serial online1998;102e3Available at:http://www.pediatrics.orgAccessed January 12, 2005Article
Cummings  PKoepsell  TRivara  FMcKnight  BMack  C Air bags and passenger fatality according to passenger age and restraint use. Epidemiology 2002;13525- 532
Kahane  C Fatality Reduction by Air Bags: Analysis of Accident Data Through Early 1996.  Washington DC National Highway Traffic Safety Administration1996;Publication HS 808 470
Durbin  DKallan  MElliott  MArbogast  KCornejo  RWinston  F Risk of injury to restrained children from passenger air bags.  Paper presented at: 46th Annual Proceedings of the Association for the Advancement for Automotive Medicine Septmber 2002 Tempe, Ariz
National Highway Traffic Safety Administration, Federal Motor Vehicle Safety Standards: Occupant Crash Protection, 62 Federal Register 12960- 129751997;
Kiuchi  T A simulation study on inflation-induced injury and New Car Assessment Program with depowered air bag.  Paper presented at: 16th International Technical Conference on the Enhanced Safety of Vehicles May 31, 1998 Windsor, Ontario
Partners for Child Passenger Safety, Interim Report.  Philadelphia, Pa The Children's Hospital of Philadelphia April 2003
Zuby  DFerguson  SCammisa  M Analysis of driver fatalities in frontal crashes of air bag equipped vehicles in 1990-98 National Automotive Sampling System, Crashworthiness Data System.  Paper presented at: Society of Automotive Engineers 2001 World Congress March 2001 Detroit, Mich
Segui-Gomez  MBaker  S Changes in injury patterns in frontal crashes: preliminary comparisons of drivers of vehicles model years 1993-1997 to drivers of vehicles 1998-2001.  Paper presented at: 46th Annual Proceedings of the Association for the Advancement for Automotive Medicine September 2002 Tempe, Ariz
Augenstein  JPerdeck  EStratton  JDigges  KSteps  J Performance of depowered air bags in real world crashes.  Paper presented at: SAE 2002 World Congress March 4, 2002 Detroit, Mich
Chidester  ARoston  T Air bag crash investigations.  Paper presented at: International Technical Conference on the Enhanced Safety of Vehicles June 4, 2001 Amsterdam, the Netherlands
Durbin  DBhatia  EHolmes  J  et al.  Partners for Child Passenger Safety: a unique child-specific crash surveillance system. Accid Anal Prev 2001;33407- 412
 The Abbreviated Injury Scale: 1990 Revision.  Des Plaines, Ill Association for the Advancement of Automotive Medicine 1990
Durbin  DRWinston  FKApplegate  SMMoll  EKHolmes  JH Development and validation of the Injury Severity Assessment Survey/Parent Report: a new injury severity assessment survey. Arch Pediatr Adolesc Med 1999;153404- 408
Korn  EGrubard  B Examples of differing weighted and unweighted estimates from a sample survey. Am Stat 1995;49291- 295
Hinch  JHollowell  WKanianthra  J  et al.  Air Bag Technology in Light Passenger Vehicles.  Washington, DC National Highway Traffic Safety Administration June27 2001;
Edwards  JSullivan  K Where are all the children seated and when are they restrained?  Paper presented at: Society of Automotive Engineers Government and Industry Meeting May 5, 1997 Washington, DC