Abstract
Motor vehicle trauma is the leading cause of death for US children aged 6–14 years. Standard sources of data on crashes involving children include police “accident” reports, hospital discharge records, statewide databases linking police reports and clinical data, and national crash injury databases. More recently trauma center based crash reconstruction methodology has been applied to pediatric populations. “Real world” crash data can be used to assess the magnitude of traffic injuries in children, to assess vehicle and restraint performance in crash situations, to identify design changes to mitigate injury, and to provide proxy data on pediatric injury tolerance.
MOTOR VEHICLE TRAUMA continues to be the leading cause of death among US children aged 6–14 years. Nearly 8 children are killed every day on US highways and another 980 are injured. Children under 15 years account for 5 percent (1,811) of all occupant fatalities and 10 percent of all persons injured in motor vehicle crashes [NHTSA 1998]. The total annual costs of pediatric motor vehicle injury for 1991 was estimated to be more than $US 7.5 billion [Miller 1995].
In addition to passenger fatalities, deaths to pedestrians and pedal cyclists comprise a significant portion of pediatric motor vehicle related trauma. Nearly one-forth of traffic fatalities to children younger than 15 are to pedestrians. Children also account for 41% of all vehicle-related deaths to pedal cyclists [NHTSA 1998].
In order to develop effective countermeasures for pediatric motor vehicle-related injury and to assess the performance of existing safety technologies, it is necessary to develop a thorough understanding of the injuries resulting from child motor vehicle trauma and of the mechanisms by which those injuries are produced. Data from “real world” crashes can help to identify problem areas in child occupant protection and may provide information needed to help assess the performance and improve the design of child restraint technologies.
DATA SOURCES FOR CRASHES INVOLVING CHILDREN
Our current understanding of child occupant protection has been built upon a combination of analysis of data from hospital records, police reports, national crash injury databases, and trauma center based crash reconstructions. It is important to understand the strengths and limitations of each data source as it pertains to child passenger safety.
POLICE REPORTS
One of the most common sources for highway safety data is police “accident” reports. While data requirements vary among the states, US police reports typically include information on vehicle type, crash configuration, restraint use, seating position, and injury. Current efforts to standardize police reports have met with limited success. Although police records have been shown to underreport both injury rates [Cercarcelli 1996, Rosman 1994, Agran 1990] and restraint use [Bents 1994, Cooper 1994], nevertheless, they have been used to estimate rates of child restraint use by age, and to estimate injury risks associated with restraint nonuse and seating position [Johnson 1994]. While police records may be useful to determine the frequencies of certain types of crash events, they cannot be used to accurately assess the magnitude of injury risks associated with crash variables because they lack sufficiently detailed injury information.
Because most police officers who file these reports have no training in crash reconstruction, the data are of limited use for engineering research. Of special concern to those doing research in child occupant protection is the large number of reports listing restraint use as “unknown” [Bents 1994]; anecdotal evidence suggests that restraint usage reported as unknown may be more common for child passengers than for adults. The accuracy of police reports has been questioned by an international survey investigating the utility of police crash data. The survey indicated that fewer than one-quarter of the respondents, 74% of which had more than 10 years experience in traffic safety, considered police “accident reports” to be a suitable source of data for either fatal or non-fatal injuries [Sivak 1988].
An important limitation of the use of police records to characterize childhood traffic injuries is selection bias of the reports themselves. Investigators in the Netherlands found, in a national survey of traffic-related injuries, that only 9% of injuries to children were documented in police records [Harris 1990]. The use of police records is also problematic for investigating injuries to pedal cyclists, where researchers have found rates of underreporting from 10% to 95% [Agran 1990, Harris 1990, Olkkonen 1990], and for investigations of non-crash, traffic-related injuries, especially pedestrian injuries to children aged 0–4 years [Agran 1990, Agran 1985].
HOSPITAL BASED CRASH RESEARCH
Hospital based research on traffic related injuries to children most often takes the form of case reports or descriptive summaries of injury patterns [Johnson D 19 Newman DK 19]. While these data are generally not population based, they have provided researchers and policy makers with information that has helped to identify numerous traffic injury risks to children and are useful in helping to classify those injury risks which bear out additional research using state or national databases.
Easily the most successful and prolific research based on hospital data is the surveillance system Dr. Agran and her colleagues in Orange County, California (1992a, 1992b, 1990a, 1990b, 1985). They developed a county-wide surveillance system that incorporated data from 9 hospital emergency departments and the county coroner’s office. By supplementing the surveillance findings with survey greatly advanced the understanding of traffic injury risks to children through their investigations of injuries to pediatric pedestrians and motor vehicle occupants.
Just as police accident reports are of limited value for research, so, too, are routine hospital records. Documentation of injuries is strongly influenced by billing procedures and protection from liability [Bents 1994]. Detailed descriptions of external soft tissue injuries, of critical importance for crash reconstruction, are notoriously incomplete in medical records. Medical records frequently do not document superficial injuries such as contusions, abrasions, and minor lacerations in sufficient detail necessary to reconstruct occupant contacts with vehicle interiors. Rather then specifying the location of each observed soft tissue injury, medical records frequently contain phrases such as “multiple abrasions and lacerations.” In addition, anthropometic data, even for such seemingly routine variables as the child’s height, are seldom recorded.
LINKED DATABASES
NHTSA has provided funds and technical support to certain states in order to foster the establishment of a Crash Outcome Data Evaluation System (CODES) that links statewide police reports with reports of pre-hospital providers of medical care, hospital discharge reports, financial data, and other medical and insurance files [Johnson 1996].
Reliable record linkage is a prerequisite for high-quality, population-based injury surveillance data [Brenner 1998]. Studies in Western Australia have documented matching of between 50 and 90% between police and hospital records [Rosman 1996] depending whether the surnames or phonetic name codes were used. While technologies for linking large data sets has undergone rapid improvement over the last decade, it is nevertheless disconcerting to note that Barancik (1985) was able to match hospital records and police reports for only 28% of injured motor vehicle passengers younger than 16.
The utility of linked databases in assessing motor vehicle injury risks to children is limited by the same factors restricting the usefulness of police records. To the extent that children, especially those injured as pedal cyclists or pedestrians, may be underreported in police records, they will be underrepresented in linked databases.
NATIONAL CRASH INJURY DATABASES
The National Highway Traffic Safety Administration maintains two national databases which contain data on children injured in motor vehicle crashes: the National Automotive Sampling Crashworthiness Data System (NASS CDS), and the Fatal Analysis Reporting System (FARS).
The FARS is a census of police-reported crashes occurring on public roads in which a fatality occurred within 30 days of the crash. FARS data are drawn from existing state records such as police reports, autopsy reports, death certificates, driver records, vehicle registrations, emergency response records and blood alcohol test reports. Some 90 elements are coded in a standardized format.
Because FARS is a census it highly accurate and easy to use to determine the magnitude of traffic safety problems with respect to fatal outcomes. However, since much of the data concerning the crash and vehicle are drawn from police reports, missing and inaccurate data can make relationships between restraint factors and outcomes more difficult to determine.
The NASS CDS collects data on a sample of police-reported, tow-away crashes stratified by crash severity; serious crashes are favored in the sampling process in order to obtain information on those crashes causing the greatest personal injury. Twenty-six field investigation teams study approximately 5000 crashes annually. Trained investigators locate and photograph the vehicles involved in selected crashes, measure vehicle damage, and interview crash survivors. Medical records are reviewed to determine the nature and severity of the crash occupants’ injuries. Injury data are coded for severity and type using the Abbreviated Injury Scale.
Both NASS and FARS data are useful in monitoring secular trends in traffic related injury. Data in the NASS CDS offer many significant improvements over previously described sources. The database contains an enormous amount of information on vehicle, safety belt, and airbag performance. And while injury data are subject to the usual caveats concerning data collected from medical records, they include the investigator’s assessment of occupant kinematics, contact points within the vehicle and injury mechanisms. Because NASS cases are drawn from stratified sample, they can be used to generate national estimates and to generate statistics about trends in the national crash experience. In spite of the vast amounts of vehicle data collected for NASS cases, however, they are of more limited use in assessing performance of child safety seats, as the data collection protocols do not allow for entering information on safety belt retractor mechanisms, safety seat failure at impact, or the presence of locking clips used with child safety seats. Until very recently, it was not possible to distinguish belt-positioning booster seats from shield boosters in the database.
SPECIAL CRASH INVESTIGATIONS
The NHTSA also conducts special crash investigations (SCI) on issues of special concern to the Agency. These case studies are not intended to form a representative sample of crash related injury, but focus on assessing the performance of recently introduced safety technologies. Using NASS data collection methodology, SCI cases have been used to collect additional information on serious injuries to children resulting from airbag deployments
In 1996, the National Transportation Safety Board (Weinstein 1998) undertook an assessment of the performance and use of child restraint systems from a sample of police-reported crashes in which at least one passenger was younger than 11 years, and in which at least one occupant was hospitalized. The NTSB devised data collection instrument that overcame many of the shortcomings of NASS with respect to collecting information on the performance of child safety seats. Medical data were abstracted from hospital records and injury severity coding was performed by clinicians experience in the assessment of motor vehicle trauma.
INSURANCE INDUSTRY DATA
By forming a consortium between academic and insurance industry researchers, Partners for Child Passenger Safety, is creating a population-based surveillance system based on a probability sample of children aged 0–15 years identified from all claims to a single vendor from 15 states and the District of Columbia. Selected families will be contacted by telephone to secure consent and will be interviewed to provide detailed information concerning restraint use at the time of crash and to report crash-related injuries. In addition, 300 cases will be selected for trauma center based reconstruction using a methodology similar to that detailed below.
This approach will provide a geographically representative crash surveillance system that is focused on pediatric traffic injuries. By collecting data on all child motor vehicle passengers meeting selection criteria, researchers will be able to compare discrepant injury/crash severity pairs, that is, severe crashes in which the child passenger was uninjured and minor crashes resulting in severe or fatal injury to the child passenger, providing data that is difficult to obtain by other methods. The interview data may be subject to reporting biases with respect to restraint use. Ongoing research is being conducted to determine the validity of parent-reported injury data.
TRAUMA CENTER BASED STUDIES
Perhaps the most important characteristic of these investigations is the interdisciplinary collaboration among clinicians, engineers, and crash reconstructionists. Many of the early advances in child occupant protection were the result of research involving isolated disciplines within the injury control sciences: epidemiologists identified injury risks, clinicians developed new diagnostic and therapeutic approaches to trauma, and engineers worked to improve vehicle and restraint designs. Researchers approached the problem in a multidisciplinary, rather than in a truly interdisciplinary manner [Winston 1996, Pless 1995].
In the early 1980s Dr. Howard Champion developed an interdisciplinary research methodology in which crash reconstructionists collaborated with medical specialists to investigate the mechanisms of traumatic injuries resulting from motor vehicle crashes. This approach was enhanced by Drs. John Siegel and Patricia Dischinger who broadened the membership of research team to include social workers, therapists, rehabilitation specialists, economists and epidemiologists, thus expanding the scope of the data collected and introducing the capacity for long-term follow-up of disabilities resulting from crash-related injuries. In 1997 the Crash Injury Research and Engineering Network (CIREN) was formed through the collaboration six trauma centers and one municipal trauma system in order to share both expertise and data and to develop a standardized approach to data collection.
Typically, trauma center based crash investigations begin with the identification of a case from hospital admission. Patients who meet eligibility criteria, based on epidemiological, clinical, or engineering interest, are then solicited for informed consent. When consent is secured, crash reconstruction teams are notified and inspections of the vehicle and scene are usually conducted within 24 hours. Crash reconstructionists review police reports, interview witnesses and emergency personnel who responded to the scene of the crash, measure vehicle deformation, and inspect the vehicle interior for evidence of occupant contact. Detailed data are collected on the use and performance of child restraints..
Prospective collection of clinical data occurs concurrently with the crash reconstruction. All injuries are documented photographically and, when appropriate, radiographically. Anthropometric measurements are taken, which in addition to the patient’s height and weight, may include seated height and measurement of upper and lower extremity lengths. The patient’s clinical course is recorded and acute care expenses are documented. Follow-up data are often collected at 6- and 12-month intervals concerning long term disability, changes in quality of life, and economic costs. The entire research team meets monthly to discuss each case with presentations of clinical and crash reconstruction data. For each injury a patient sustains, the researcher team determines the most likely source and mechanism of injury.
This approach provides the most accurate and detailed medical detail available for crash-related injuries. Like the SCI cases, investigators can concentrate data collection efforts on documenting the performance of emerging restraint technologies. In addition, many centers have followed patients 6- and 12-months post injury to provide new insights on long-term outcomes of motor vehicle related injuries. By providing more detailed information on restraint usage and injury outcomes, data gathered from trauma center based studies provides a complement to the data extracted from population-based databases such as NASS and FARS.
It is important to recognize that trauma center based reconstruction research is not population based and cannot be used to make national estimates of injury risk. Furthermore, like the large national databases, data are not collected on “saves,” that is, severe crashes in which child passengers were uninjured. The high cost per case investigation (> $US 3,000) limits the broader application of this highly detailed methodology.
The application of interdisciplinary crash reconstruction research to child occupant protection issues has revealed several challenges that are especially problematic for investigating injury to pediatric populations. The two most important issues in this regard are in securing informed consent/assent and assessing restraint usage.
Obtaining informed consent can be more time consuming for pediatric patients, as custodial authority may be shared between parents who have little direct communication with each other. Additionally, if one or both parents sustains incapacitating injuries in the same crash, non-injured parents or relatives may be reluctant to sign consent documents. It is further suspected that feelings of parental guilt relating to injury causation or restraint use may further impair efforts to obtain consent in a timely manner. Delays in obtaining consent, and hence in vehicle inspection, can result in the degradation of forensic evidence of occupant contact and make biomechanical assessment more difficult.
Accurate assessment of restraint use for child passengers is made more difficult by the more limited nature of relevant forensic data, the greater variety of available restraints, and the increased likelihood of restraint misuse. The child’s smaller body mass is less likely to stretch belt webbing or result in other indicators of belt loading typically seen in adult occupants. Well-meaning bystanders may remove a crying child from a safety seat prior to the arrival of emergency responders. Drivers may misrepresent a child’s restraint status at the time of a crash due to guilt or fear of litigation. Because of the potentially conflicting nature of data concerning restraint use, the researcher is compelled to seek data from a wide variety of sources and to make assessments based on a preponderance of evidence.
LESSONS FROM REAL WORLD CRASH DATA
The data collection and research methodologies have produced a wealth of information on traffic related injuries to children. Using surveillance based on the cooperation of nine emergency rooms and the local coroner’s office, Dr. Agran and her colleagues have been able to demonstrate age-related differences in safety belt effectiveness in children (1992a), identify circumstances associated with on-lap travel for infants (1992b), document injury risks to children riding in the pack of pick-up trucks (1990), and to describe injuries to children caused by noncrash falls and ejections (1985).
Analysis of statewide CODES databases has begun to yield useful information on childhood motor vehicle trauma. By electronically linking clinical and police records, these databases have been used to establish local estimates of injuries associated with restraint use and seating position in children and to assess differences in restraint efficacy by age [Corneli 1998, Cook 1998]. CODES data have been used by safety advocates working to enact state laws promoting pediatric traffic safety issues including primary enforcement of child restraint laws and graduated licensing [Nechodom 1998].
Many of the findings initially reported by hospital based studies have been further confirmed through the use of the national databases. Braver (1998), using eight years worth of FARS data clearly demonstrated the benefits, with respect to mortality risks, of the rear seating positions and of the rear center versus outboard seating position. Studies by Huelke (1995) and Johnson (1995) have resulted in similar findings for injury outcomes using NASS data.
Data from NHTSA’s Special Crash Investigations and the NTSB study on child restraint effectiveness have played a major role in documenting the biomechanics of passenger airbag related fatalities to children in rearward facing infant safety seats and to unrestrained children. This research was influential in helping to introduce changes in US federal regulations that have allowed “depowered” airbags and contributed also to new rulemaking regarding the installation of “on/off” switches for passenger air bags.
Interdisciplinary crash reconstruction methods have only recently been applied to pediatric populations, but have already resulted in findings that have direct implications for reducing traffic injuries in children. Winston, et al., (1998) investigated bicycle-related injuries to children and determined that crash avoidance maneuvers, often performed by inexperienced riders, frequently caused abdominal “impalement” by the handlebars resulting in pancreatic injury. The authors have suggested these injuries could be mitigated by changes in handlebar design.
Another application of crash reconstruction research has been to document injury patterns and costs associated with specific types of safety seat misuse. [Gotschall 1998, Weinstein 1998]. While observational studies can give estimates of the prevalence of specific types of safety seat misuse, information linking misuse with injury rates and injury severity provides additional information that can be useful in targeting limited resources for child occupant protection.
New restraint technologies are continually being introduced. Unlike the clinical milieu, where the benefit and safety of all new medical technologies and treatments must be rigorously demonstrated to the Federal Drug Administration, the principle requirement of manufacturers of automotive restraint devices is to demonstrate that new technologies do not violate existing safety regulation. For example manufacturers of devices that modify the fit of safety belts, marketed primarily to parents of young children, need not demonstrate that they improve the performance of belt systems, rather, they must merely demonstrate that belt performance is not degraded below existing safety standards. As new child restraint technologies are developed manufacturers and safety advocates will continue to rely on “real crash” data to more fully evaluate performance. Researchers should not look to any single methodology or data source for definitive assessment of restraint effectiveness but must continue to collaborate using complementary techniques to help improve understanding of childhood traffic injury.
LITERATURE CITED
- Agran PF, Castillo DN, Winn DG. Comparison of motor vehicle occupant injuries in restrained and unrestrained 4- to 14-year olds. Accid Anal & Prev. 1992a;24:349–355. doi: 10.1016/0001-4575(92)90047-m. [DOI] [PubMed] [Google Scholar]
- Agran PF, Winn DG, Castillo DN. On-lap travel: still a problem in motor vehicle crashes. Pediatrics. 1992b;90:27–29. [PubMed] [Google Scholar]
- Agran PF, Castillo DN, Winn DG. Limitations of data compiled from police reports on pediatric pedestrian and bicycle motor vehicle events. Accid Anal & Prev. 1990a;22:361–370. doi: 10.1016/0001-4575(90)90051-l. [DOI] [PubMed] [Google Scholar]
- Agran PF, Winn DG, Castillo DN. Pediatric injuries in the back of pickup trucks. JAMA. 1990b;264:712–716. [PubMed] [Google Scholar]
- Agran PF, Dunkle DE. A comparison of reported and unreported noncrash events. Accid Anal & Prev. 1985;17:7–13. doi: 10.1016/0001-4575(85)90003-x. [DOI] [PubMed] [Google Scholar]
- Barancik JI. Fife: Discrepancies in vehicular crash injury reporting: Northeastern Ohio Trauma Study, IV. Accid Anal & Prev 1985. 1985;17:147–154. doi: 10.1016/0001-4575(85)90017-x. [DOI] [PubMed] [Google Scholar]
- Bents FD, Dischinger PC, Siegel JH. Trauma center based crash investigation research: methodologies and applications. Society of Automotive Engineers (SP-1042); Warrendale, PA. 1994. [Google Scholar]
- Brenner H, Schmidtmann I. Effects of record linkage errors on disease registration. Methods Inf Mad. 1998;37:69–74. [PubMed] [Google Scholar]
- Buchbinder R, Goel V, Bombardier C. Lack of concordance between the ICD-9 classification of soft tissue disorders of the neck and upper limb and chart review diagnosis: one steel mill’s experience. Am J Ind Med. 1996;29:171–182. doi: 10.1002/(SICI)1097-0274(199602)29:2<171::AID-AJIM7>3.0.CO;2-V. [DOI] [PubMed] [Google Scholar]
- Cercarelli LR, Rosman DL, Ryan GA. Comparison of accident and emergency with police road injury data. J Trauma. 1996;40:805–809. doi: 10.1097/00005373-199605000-00021. [DOI] [PubMed] [Google Scholar]
- Cook L, Knight S, Nechodom P, Dean JM. The effect of seating position on injuries to children in motor vehicle crashes. National Congress on Childhood Emergencies Syllabus, Health Resources & Services Administration; Rockville, MD: 1998. [Google Scholar]
- Cooper PJ. Estimating overinvolvement of seat belt nonwearers in crashes and the effect of lap\shoulder restraint use on different crash severity consequences. Accid Anal & Prev 1994. 1994;26:263–275. doi: 10.1016/0001-4575(94)90096-5. [DOI] [PubMed] [Google Scholar]
- Corneli HM, Reading JC, Nechodom P, Dean JM. Decreased protection with increasing age: Problems affecting the efficacy of motor vehicle safety restraints for children. National Congress on Childhood Emergencies Syllabus, Health Resources & Services Administration; Rockville, MD: 1998. [Google Scholar]
- Gotschall CS, Eichelberger MR, Morrissey JR, Better AI, Reardon J. Injury patterns associated with child restraint misuse. Child Occupant Protection 2nd Symposium Proceedings; Warrendale, PA: Society of Automotive Engineers; 1998. pp. 187–194. [Google Scholar]
- Miller T. Childhood Injury: Cost & Prevention. National Public Services Research Institute; Landover, MD: 1996. [Google Scholar]
- Harris S. The real number of road traffic accident casualties in The Netherlands: a year-long survey. Accid Anal & Prev. 1990;22:371–378. doi: 10.1016/0001-4575(90)90052-m. [DOI] [PubMed] [Google Scholar]
- Huelke D, Compton CP. The effects of seat belts on injury severity of front and rear seat occupants in the same frontal crash. Accid Anal & Prev. 1995;27:835–838. doi: 10.1016/0001-4575(95)00021-6. [DOI] [PubMed] [Google Scholar]
- Johnson SW, Walker J. NHTSA Technical Report: The Crash Outcome Evaluation System. Washington, DC: Department of Transportation, National Highway Traffic Safety Administration; Jan, 1996. DOT HS 808 338. [Google Scholar]
- Johnston C, Rivara FP, Soderberg R. Children in car crashes: analysis of data for injury and use of restraints. Pediatrics. 1994;93:960–965. [PubMed] [Google Scholar]
- Johnson DL, Falci S. The diagnosis and treatment of pediatric lumbar spine injuries caused by rear seat lap belts. Neurosurg. 1990;26(3):434–441. doi: 10.1097/00006123-199003000-00010. [DOI] [PubMed] [Google Scholar]
- Morrison P, Kjellstrom T. A comparison of hospital admission data and official government statistics of serious traffic accident injuries. N Z Med J. 1987;100:517–520. [PubMed] [Google Scholar]
- National Highway Traffic Safety Administration: Traffic Safety Facts – 1996 Children. National Center for Statistics & Analysis; Washington, DC: 1998. [Google Scholar]
- Nechodom PJ, Knight S, Dean JM. National Congress on Childhood Emergencies Syllabus, Health Resources & Services Administration; Rockville, MD: 1998. Graduated licensing: A sensible solution to a growing problem. [Google Scholar]
- Newman KD, Bowman LM, Eichelberger MR, Gotschall CS, Taylor GA, Johnson DL, Thomas M. The lap belt complex: Intestinal and lumbar spine injury in children. J Trauma. 1990;30(9):1133–1140. [PubMed] [Google Scholar]
- Olkkonen S, Lahdenranta U, Tolonen J, Slatis P, Honkanen R. Incidence and characteristics of bicycle injuries by source of information. Acta Chir Scand. 1990;156:131–136. [PubMed] [Google Scholar]
- Pless IB. Interdisciplinary and multidisciplinary are not synonymous. Inj Prevent. 1995;1:65. doi: 10.1136/ip.1.2.65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rosman DL, Knuiman MW. A comparison of hospital and police road injury data: Sivak M, O’Day J: Survey concerning international computer files of road accidents. Accid Anal & Prev. 1988;20:459–463. doi: 10.1016/0001-4575(88)90044-9. [DOI] [PubMed] [Google Scholar]
- Rosman DL. The feasibility of linking hospital and police road crash casualty records without names. Accid Anal & Prev. 1996;28:271–274. doi: 10.1016/0001-4575(95)00058-5. [DOI] [PubMed] [Google Scholar]
- Siegel JH, Mason-Gonzales S, Dischinger P, et al. Safety belt restraints and compartment intrusions in frontal and lateral motor vehicle crashes: Mechanisms of injuries, complications, and acute care. J Trauma. 1993;37:736–748. doi: 10.1097/00005373-199305000-00017. [DOI] [PubMed] [Google Scholar]
- Sivak M, O’Day J. Survey concerning international computer files of road accidents. Accid Anal & Prev. 1988;20:459–463. doi: 10.1016/0001-4575(88)90044-9. [DOI] [PubMed] [Google Scholar]
- Weinstein EB, Sweeney MM, Barber M, Eastwood MD, Osterman JG, Roberts JV. The effect of size appropriate and proper restraint use on injury severity in children. Child Occupant Protection 2nd Symposium Proceedings; Warrendale, PA: Society of Automotive Engineers; 1998. pp. 181–186. [Google Scholar]
- Winston FK, Shaw K, Kreshak AA, Schwarz DF, Gallagher PR, Cnaan A. Hidden Spears: Handlebars as injury hazards to children. Pediatrics. 1998 Sep; doi: 10.1542/peds.102.3.596. in press. [DOI] [PubMed] [Google Scholar]
- Winston FK, Schwarz DF, Baker SP. Biomechanical epidemiology: A new approach to injury control research. J Trauma. 1996;40:820–824. doi: 10.1097/00005373-199605000-00024. [DOI] [PubMed] [Google Scholar]
- Winston FK, Reed R. Air Bags and Children: Results of a National Highway Traffic Safety Administration Special Investigation into Actual Crashes. Proceedings of the 40th Stapp Car Crash Conference; 1996. pp. 383–389. [Google Scholar]