views updated


Accidents, according to dictionary definitions, are events that happen by chance and are not predictable and therefore are not preventable. In contrast, from a public health perspective, injuries were first clearly conceptualized by William Haddon (1964) as damage done to the body as a result of often predictable and therefore preventable energy exchange. This energy exchange may be kinetic, thermal, or chemical. For example, kinetic energy can result in injury when someone falls, hits the dashboard in a car crash, is penetrated by a bullet, or is hit on the head with a baseball bat. Injuries associated with thermal energy result in burns. Chemical energy can create injury through contact with caustics or via ingestion of a wide assortment of poisonous substances. Injuries can also occur when there is a lack of energy (e.g. frostbite) or an essential agent such as oxygen (e.g. drowning). Consequently, because of the connotation that accidents happen by chance, and therefore aren't preventable, injury control specialists no longer use the term "accident," preferring to refer to injury occurrences in terms of the specific circumstances— "crashes," "falls," "fires"—or as "events" or "incidents."

The literature on traumatic injury includes both unintentional injuries (known in common parlance as "accidents") and intentional injury (i.e. violence). Some types of injury are not so easily categorized in this way—for example, traumatic brain injury resulting from having been a shaken baby. While this type of injury is inflicted, it does not usually result from a typical violent act, which involves intent to do harm. Rather, it often results from attempts to quiet a cranky baby by an individual unaware of the risks. Yet the physiological results are often devastating.

Injuries occur as part of a process that can be understood within the social-ecological framework as described by Urie Bronfenbrenner in 1979. This framework can be extended in a public health context to conceptualize behaviors or health outcomes such as injury as a result of interactions among individual characteristics (intrapersonal level) and interactions among individuals (interpersonal), as well as with the physical and social environment, including institutional and cultural elements. Developmental characteristics enhance or reduce the possibilities that a child will experience an injury. For example, young children are not developmentally prepared cognitively to assess and avoid risk. Toddlers are curious and developmentally eager to explore. They also have relatively large heads that can throw off their balance and contribute to their risk of falling headfirst (e.g., into a toilet or bucket) while exploring their surroundings. Similarly, characteristics of young children learning to pull themselves up may create problems as they may pull objects down on themselves (e.g. a hot cup of coffee or books).

Incidence of Injuries to Children

To fully understand the problem of childhood injuries, it is important to examine how many cases of injury are occurring, in which population groups, for example, distributions by age, race, and gender. Data across sources are not always comparable due to differences in measurement, definitions, or data compilation practices. Hence, understanding of the incidence of any public health problem necessitates careful review of the data sources and consideration of what they can and can't reveal about the problem.

Measurement Issues

As with any public health problem, it is important to have data available from routine surveillance—the systematic, ongoing collection of information about illnesses, injuries, or events. To fully understand the magnitude of the problem, one must consider both the mortality (fatalities) and morbidity (nonfatalities) associated with injury. Mortality is readily assessed using data available from death certificates. Data are coded as to the underlying causes of death and can yield valuable information about factors contributing to the injury. In states having medical examiner systems, these records can provide more detail about the circumstances of death.

The Centers for Disease Control and Prevention report that for every child who dies as a result of an injury, there are an estimated 34 who are injured seriously enough to require hospitalization, and another 1,000 who receive medical care in an outpatient setting. Even more children experience injuries that result in limitations of their activities or cause discomfort but that do not require medical attention. Hospitalized injuries can be examined using hospital discharge data compiled in each state as well as nationally. The standards of reporting, however, are not as stringent as with death statistics. Underlying causes are not uniformly coded in all states, some requiring that this information be included and others reporting on a voluntary basis.

Data about nonhospitalized cases are even less uniformly compiled. There are few statewide and no national reporting systems for emergency department cases. Injuries treated in private doctor's offices, urgent care centers, or by school nurses may or may not be routinely documented, but there are no national record-keeping systems to monitor these events systematically. Consequently, information about injury morbidity is of variable quality.

Types of Injuries

Causes of injury death differ by age and by type of injury, as demonstrated in Table 1. The data presented in these figures reflect only mortality.

In 1998, 18,292 U.S. children died as a result of injuries, for an overall death rate of 23.55 per 100,000 population. Of these fatalities, 12,416 were classified as unintentional, 3,461 as homicide, and 2,061 as suicide. Rates varied by race and gender, with boys (aged zero to nineteen) experiencing an overall death rate of 32.04 per 100,000, compared to 14.63 for girls. As shown in Table 2, racial differences reveal higher rates for minority youth, with Native Americans experiencing the highest rates of unintentional injury and suicide while African Americans' rates for homicide exceed those of other ethnic groups. Some of this difference is attributable to environmental conditions associated with poverty (African Americans and Native Americans) and/or living in rural areas (Native Americans) where risks are high and rapid access to medical care is less readily available.

Injury patterns differ by both age and type of event, as demonstrated in Tables 1 and 3. For all age groups, motor vehicle crashes were a significant cause of fatalities in 1998. The youngest children (infants) are most likely to be fatally injured as a result of suffocation, motor vehicle traffic, and drowning/submersion. For toddlers, the leading causes of injury death are motor vehicle traffic, drowning/submersion, and homicide, while for those in the five to fourteen age group (elementary and middle-school age), most fatal injuries result from motor vehicle traffic, homicide, and drowning/submersion. In contrast, teenagers are fatally injured as a result of motor vehicle crashes, homicide, and suicide.

As shown in Table 3, various causes of injury fatality exhibit differing patterns depending on age. In part, this is a function of different developmental factors that impinge on children's abilities to avoid injury events (e.g., being able to walk without falling, being able to make judgments about avoiding risks, testing authority, having transient depression, acting impulsively). In addition, children have greater or lesser exposures to different risky situations (e.g., crossing streets without a parent, playing sports, riding a tricycle near traffic, being a teenage driver, owning or carrying firearms, working on a construction crew) based on the practices of their parents (e.g., level of supervision provided, drinking and smoking behaviors, disciplinary practices); or because of social, economic, or cultural factors (e.g., access to affordable, high-quality child care; well-constructed homes that are equipped with fire safety devices such as smoke detectors; practices of riding in the back of pickup trucks; use of hazardous products such as firearms, baby walkers, in-line skates, or trampolines.


Morbidity (i.e., nonfatal injury) is both much more common than mortality and presents different patterns of injury. In 1997, 213,000 hospitalizations of children ages zero to fifteen in the United States resulted from injuries, while in 1996 alone an estimated 8.71 million in this age group were treated and released from emergency care. Using 1990s data from North Carolina on unintentional injury for children under age fifteen, Table 4 lists the differences in injury patterns by age group for the leading causes of death, of nonfatal, hospitalized injuries, and estimated incidence of injuries treated in outpatient settings. This demonstrates the importance of considering the range of injury outcomes in establishing priorities and devising preventive strategies.

Furthermore, several child and adolescent injury problems are not easily recognized in the health care system at all, yet have profound effects on injury occurrence and developmental outcomes. One type of event is child abuse and neglect. Reporting through social services systems and national surveys indicates that as many as 1.5 million children a year experience physical or sexual abuse at the hands of an adult caretaker. Likewise, children and adolescents may be sexually assaulted or raped by strangers or experience injuries in dating relationships, similar to domestic violence experienced by adults. National estimates indicate that as many as 1.4 million persons under age eighteen experience sexual assault or rape each year. In addition, it is estimated that 8.8 percent of adolescents may experience violence in their dating relationships, including incidents of hitting, slapping, or being physically hurt.

Other types of circumstances that put teenagers at risk of nonfatal injuries are sports participation and employment. Together, these two categories of injury result in more emergency care for teenagers than any other categories. Sports and occupational injury are the two most common categories. National data suggest that as many as 3.5 million youth under age 14 are injured in sports and recreational activities annually, 775,000 of them requiring emergency care. Seventy-five percent are boys. Similarly, youth who are employed are exposed to a variety of hazards, with as many as half experiencing injuries at work.

Developmental Delays as a Result of Injury

Just as development affects whether a child is at risk of experiencing an injury-producing event, so is development affected by being injured. Injuries with the greatest impact in the behavioral and emotional development of the child are head injuries, specifically traumatic brain injuries, and severe burns. Little research has been done on the outcome of other types of injuries in children and adolescents.

Brain Injury

As reported in 1996, more than 100,000 children and adolescents are hospitalized each year with traumatic brain injury (TBI), of which about 88 percent are classified as mild, 7 percent moderate, and 5 percent severe. Although children under five years of age have the lowest rate of TBI, the severity of injury is disproportionately high. Evaluations of neurological and psychological outcomes following pediatric TBI, however, have been focused almost exclusively on school-age children and adolescents. The few studies that have assessed the consequences of mild TBI for infants and/or preschoolers identified lower IQ scores in infants and young children than in older children who suffered a mild TBI. Philip Wrightson, Valerie McGinn, and Dorothy Gronwall (1995) studied pre-schoolers who sustained mild head injuries and reported that deficits in cognitive development, identified at twelve months after injury, were significantly associated with reading ability at age six and a half. It is not clear if mild TBI and concussion result in long-term problems in children given the difficulty in differentiating effects that may be a direct consequence of mild TBI or of learning disabilities or attention deficits in the TBI group, which do not manifest themselves until later in development.

Follow-up studies of severe TBI in school-age children and adolescents identified physical, cognitive (e.g., lower IQ scores), language, and psychological deficits that may be temporary or permanent. The most common language problems are word retrieval difficulties and decreased speed of information processing. Follow-up studies for as long as three years after the TBI reveal persistent neurobehavioral deficits. Acute behavioral problems include emotional outbursts, restlessness, and low tolerance to stimulation.

Cognitive delays have significant effects on social, educational, and vocational prognosis of children following TBI. For example, in a study in which TBI children were compared with age, gender, and grade-matched peers, Kenneth Jaffe and his colleagues found that more severe head injury was associated with lowered neurobehavioral functioning. Developmental deficits include lack of attention and long, heightened distractibility, short-term memory impairment, difficulty with logical thinking and reasoning, slowed reaction time, and impaired spatial or visual motor skills. The duration of post-traumatic amnesia was reported to be the best predictor of these deficits. Severe TBI results in significant neurobehavioral deficits that are persistent after the first year and that are related to the severity of the head injury.

The effects of age at injury on subsequent cognitive development are unclear. In studies examining cognitive, social, academic, and vocational outcomes following severe TBI, some have reported that age at injury was not predictive of long-term outcome. Others identified more severe consequences in children ages zero to six than in older children. Several studies of outcomes following TBI failed to find associations between age at injury and either severity of cognitive consequences and the rate of recovery of neuropsychological, or behavioral disturbance. Several other studies, however, suggest that trauma to the brain may have more severe outcomes and more significant long-term consequences for infants and preschoolers than for older children. A number of findings regarding the language and memory effects of TBI in infants, school-age children, and adolescents are also consistent with the hypothesis that skills in a rapid stage of development at the time of TBI were more adversely affected than already well-established skills.

A study that determined the long-term outcome of severe brain injury in preschoolers revealed that none of the children who were younger than four years old at the time of the severe brain injury were able to live independently or work full-time in adulthood. Other studies also reported abnormalities following TBI in 40 percent of children with minimal or no loss of consciousness.

Injury can have important consequences for family functioning, which in turn adversely affects the family's ability to assist the child in achieving her full potential. Most of these adverse effects are seen in families with children who sustained severe head injury. Several studies have revealed that the level of family functioning before the injury was a better predictor of family functioning at one year after the injury than the severity of the TBI. The researchers inferred that family dysfunction prior to the injury event was likely to continue following the injury. Parental overreaction and family dysfunction were likely to exacerbate the child's emotional reaction to the injury. Poverty can exacerbate these problems and further decrease the level of functioning of children following TBI.


Burns also produce devastating results for child development. The effects of the injury on the psychological and emotional health of children and the family demonstrate inconsistent findings.

Studies conducted in the 1970s revealed that 50 percent of children showed signs of emotional disturbance several years after their injury. Studies in the 1980s indicated that the prevalence of these symptoms might be much lower and that the level of family and child functioning prior to the injury were the most important predictors of postburn adjustment and coping. A survey of burned children in 1985 found psychosocial maladjustment in 15 percent of the children. For some of the children and families, however, psychological disturbance existed before the injury and might have hindered emotional and even physical recovery afterward. The premise that visible scarring is more damaging than hidden burn scars has not been studied well. Adaptation takes place over a period of years. In the first couple of years after the burn, children have developmental regression, phobias, and various other symptoms. These symptoms progressively resolve in the subsequent years.

Parents are frequently stressed by the child's behavior, depressed over the burn and the child's future, and have feelings of guilt regarding the circumstances leading to the injury. Some research has shown that parental stress clearly differentiated parents of burned children from other parents; children with better psychological adjustment come from families that have higher levels of cohesion, independence, and more open expressiveness. This has important implications for interventions as mothers and burned children can be identified and given special support at the time of the initial care for the child's injury.

Strategies to Ameliorate Injuries

Drawing on the same social-ecologic principles that help describe the factors affecting the occurrence of injuries, William Haddon developed two frameworks for conceptualizing injury interventions. The first outlined injury interventions aimed at different phases in the energy transfer process. The ten countermeasures were directed at reducing the potential for damage to the body from an energy source. Table 5 shows examples pertaining to reducing childhood bicycle crash injuries.

Later, Haddon proposed a different, but similar, model to help stimulate intervention development: the Haddon Matrix. This model is widely used to identify potential interventions, and has two dimensions. The first depicts the target of change, including the person at risk (or the person's caretaker); the source of energy (e.g., products to which children are exposed—toys, motor vehicles, household poisons); the physical environment (e.g., playgrounds, roadways, homes, schools, day-care facilities); or the social environment (e.g., norms about behaviors such as drinking or child discipline; regulatory policies governing drinking ages, playground safety, or use of bicycle helmets). The second dimension depicts the phase at which the intervention has its effect— whether at the time of the injury event (e.g., a seat belt that deploys during a crash); at a time prior to the event, helping to prevent the event (e.g., antilock brakes); or after the event to reduce the effects of the event (e.g., gas tanks engineered to prevent explosions). Putting these two dimensions together creates a four by three matrix. Using the matrix as a brainstorming tool, one can fill in the twelve cells with ideas of potential interventions (see example in Table 5 pertaining to bicycle safety).

Choosing among the multiple alternative intervention ideas can be complex but several principles can help sort out the types of intervention ideas. Interventions may be voluntary or mandatory. And they might be active or passive. Active interventions require that the person to be protected takes protective action (e.g., putting on a bike helmet, testing water temperature in the baby's bathtub). In contrast, passive interventions are those that do not require action on the part of the individual being protected or deliberate action by a caretaker. Rather, they are engineered or set up such that protection is automatic (e.g., installation of airbags in cars, setting the temperature for hot water heaters at a safe level; automatic shutoffs on appliances). In general, interventions that require less individual effort result in more success. That is, passive interventions tend to be more successful than ones that require individual action. There are significant trade-offs, however, in making choices among the different options.

A third dimension added to the Haddon Matrix by Carol Runyan addresses the kinds of trade-offs that frequently occur in public health decision making. By considering alternative interventions derived from the two-dimensional matrix, a planner can examine the value issues involved in each choice and compare various options. As a result, decisions can be made more logically and the rationale for a given decision can be explained to others. For example, an intervention requiring restrictions on the purchase of firearms limits the freedom of gun enthusiasts, while at the same time affording greater freedom for children to live in environments with fewer guns and thus increasing child safety. Likewise, some interventions are more costly than others, requiring design changes in products (e.g., improvements in car design to reduce rollovers in crashes) or in the design of physical space (e.g., changing the equipment on a playground at a child-care facility). Other strategies require that funds be made available to make special provisions for protecting children without personal resources (e.g., car seat or smoke detector give-away programs for low-income families), while limiting higher-income families' access to the funds.


Injury is a major public health problem for children. Data to assess the magnitude of the problem are of variable quality and in need of continuous improvement, but it is clear that special risks exist for certain subgroups of the population, including different risks by ethnic group, age, and gender. It is important to examine data from multiple sources to understand the problem, including information about deaths as well as injuries resulting in hospitalization or outpatient (e.g., emergency department) care. Other types of injuries are not as easily captured in these data systems and require efforts to estimate risks through other means—for example, collecting data from social services or through research surveys about child abuse, from schools about sports injury, or from youth about employment-related injuries.

Research tracking the long-term effects of different types of injury suggests that the developmental outcomes of more severe injuries may be profound.

Injuries are not accidents. Causes can be identified and preventive strategies developed. There are numerous interventions to choose from in addressing prevention of traumatic injuries, each with its own advantages and disadvantages.



Bronfenbrenner, Urie. The Ecology of Human Development. Cambridge, MA: Harvard University Press, 1979.

Dunn, Kathleen, Carol Runyan, Lisa Cohen, and Michael Schulman. "Teens at Work: A Statewide Study of Jobs, Hazards, and Injuries." Journal of Adolescent Health 22, no. 1 (1998):19-25.

Finkelhor, David, Gerald Hotaling, I. A. Lewis, and Christine Smith. "Sexual Abuse in a National Survey of Adult Men and Women: Prevalence, Characteristics, and Risk Factors." Child Abuse and Neglect 14, no. 1 (1990):19-28.

Haddon, William, Edward Suchman, and David Klein. Accident Research: Methods and Approaches. New York: Harper and Row, 1964.

Hennes, Halim, Martha Lee, Douglas Smith, John R. Sty, and Joseph Losek. "Clinical Predictors of Severe Head Trauma in Children."American Journal of Diseases of Children 142 (1988):1045-1047.

Jaffe, Kenneth, M., Gayle C. Fay, Nayak Lincoln Polissar, KathleenM. Martin, Hillary A. Shurtleff, J'May B. Rivara, and Richard Winn. "Severity of Pediatric Traumatic Brain Injury and Neurobehavioral Recovery at One Year: A Cohort Study." Archives of Physical Medicine and Rehabilitation 74 (1993):587-595.

Koskiniemi, Marjaleena, Timo Kyykka, Taina Nybo, and Leo Jarho. "Long Term Outcome after Severe Burn Injury in Pre-schoolers Is Worse than Expected." Archives of Pediatric Adolescent Medicine 149 (1995):249-254.

Runyan, Carol. "Using the Haddon Matrix: Introducing the Third Dimension." Injury Prevention 4 (1998):302-307.

System YRBS. "Youth Risk Behavior Surveillance: United States, 1999." Mortality and Morbidity Weekly Report 49 (2000):1-96.

Theodore, Andrea, and Desmond Runyan. "A Medical Research Agenda for Child Maltreatment: Negotiating the Next Steps." Pediatrics 104 (1999):168-177.

Wrightson, Philip, Valerie McGinn, and Dorothy Gronwall. "Mild Head Injury in Preschool Children: Evidence That It Can Be Associated with a Persisting Cognitive Defect." Journal of Neurology, Neurosurgery, and Psychiatry 59 (1995):375-380.

Carol W.Runyan

Janet Abboud DalSanto

Kristen L.Kucera