Objective To describe risk factors for injury death among infants in the United States by the specific external cause of death.
Methods. Data were analyzed from the US-linked birth/infant death files for the years 1983–1991. Potential risk factors for injury death were identified from birth certificate data and included both maternal and infant factors. Injury rates were calculated by external cause of death. Characteristics of infants who died from an injury were compared with those of the entire birth cohort. The independent effect of potential risk factors was assessed in multivariate analyses using a case–control study design.
Results. A total of 10 370 injury deaths were identified over the 9-year study period (29.72/100 000 live births). The leading causes of death were homicide, suffocation, motor vehicle crashes, and choking (inhalation of food or objects). There was no significant temporal trend in the overall rate of injury death; however, this was because significant increases in the rates of death from homicide (6.4%/year) and mechanical suffocation (3.7%/year) were offset by decreases in rates of death from fires (−4.7%/year) and choking (−4.6%/year). In adjusted analyses, infants born to mothers with no prenatal care, <12 years of education, two or more previous live births, Native American race, or <20 years of age were at twice the risk of injury death compared with the lowest risk groups (initiation of prenatal care in the first trimester, ≥16 years of education, no previous live births, white, or ≥25 years of age). When analyzed by the specific cause of death, the factors that were associated most strongly with death varied. For example, Native Americans were at greatest risk of a motor vehicle related death (compared with whites: OR: 3.6; 95% CI: 1.8–7.1), and infants with birth weights of <1500 g were at greatest risk of death attributable to inhalation of food (compared with ≥2500 g: OR: 9.6; 95% CI: 3.3–28.0) or objects (OR: 11.8; 95% CI: 4.5–30.5).
Conclusion. A number of sociodemographic characteristics are associated with an increased risk of injury-related death in infants. The strength of associations between specific risk factors and death varies with the external cause of death, thus identifying high-risk subgroups for targeting of cause-specific interventions and simultaneously increasing our understanding of the individual and societal mechanisms underlying these tragedies. infant, injury, suffocation, motor vehicle, homicide, drowning, inhalation, fire.
In the United States, injuries are the leading cause of death among persons between the ages of 1 and 44 years.1Although injuries account for <5% of deaths in the first year of life, they are the third leading cause of death during the postneonatal period, exceeded in number only by deaths caused by congenital malformations and the sudden infant death syndrome.1Although researchers have long sought an overarching conceptual framework that would allow the development of interventions that would address multiple causes of injuries, primary prevention efforts are largely cause-specific, such as the use of infant car seats to prevent injuries sustained in motor vehicle crashes.2 ,3
Previous national studies on fatal injuries in infants have relied primarily on information available on death certificates and therefore have been unable to examine the association between maternal, sociodemographic, and birth characteristics and injury death. However, regional studies have identified associations between maternal factors, such as young maternal age or high parity, and the risk of death from an intentional or unintentional injury in infancy.4–10Many of these studies have either grouped all injuries together or dichotomized them to intentional and unintentional groupings. Yet, risk profiles probably vary by cause of death, for example the highest risk group for death attributable to a motor vehicle crash is likely to differ from the highest risk group for death attributable to a residential fire. This information is crucial for appropriate targeting of cause-specific interventions.
The purpose of this study is to identify risk factors for injury death in infants <1 year of age using data from the National Center for Health Statistics linked infant birth/death certificate datasets. By using this large, national dataset we are able to provide data on injury rates and risk factors for injury death by the specific external cause of death.
Data were analyzed from the national linked birth/infant death datasets for the years from 1983–1991.11 These datasets include birth certificate information for all live births occurring in the United States as well as death certificate information for all deaths that occur within the first year of life linked to that individual's birth certificate information. Injury deaths were identified as deaths for which the underlying cause of death was coded as an injury on the death certificate (external cause of injury codes [E codes] E800 to E999 in the 9th revision of the World Health Organization's International Statistical Classification of Diseases).12 Deaths caused by surgical or medical procedures or misadventures (E870–E879) and deaths attributable to adverse effects of medicine and biologics in therapeutic use (E930–E949) were excluded. Deaths caused by late effects of a particular injury were classified with the original injury; for example, a death attributable to late effects of accidental poisoning would be included with poisonings.
Potential risk factors for injury death were identified from birth certificate information and included maternal factors (age, education, race, trimester of first prenatal visit, number of previous live births, and marital status) and infant factors (gender, birth weight, and gestational age). Cases were first compared with the entire birth cohort. Overall rates and cause-specific injury rates were calculated as deaths per 100 000 live births and reflect true rates in the population. Crude relative risks were calculated with data from the entire birth cohort. Significance testing was not performed because there were no statistical adjustments and the entire population was represented. Overall trends and cause-specific trends in injury rates were analyzed with Poisson regression techniques using the SAS statistical software logistic procedure.13
One control was selected randomly for each eligible case. Cases and controls were matched on year of birth. Crude ORs were determined for each potential risk factor. To determine the independent contribution of maternal and infant factors, those factors that were statistically significant (χ2; P < .05) in bivariate analyses were entered into multiple logistic regression models.14 All variables were entered initially into the model, as each variable was significant in bivariate analyses. For all injuries combined, each risk factor remained significant after adjusting for all others. For some specific causes of injury a single risk factor (such as gender in the case of motor vehicle-related injuries) was not statistically significant; however, these instances were rare and thus for consistency all variables are retained in the final models. Missing data were reported for at least one variable in 21% of cases and 19% of controls. The majority of missing responses were for maternal education (missing data for 18% of cases and 17% of controls) which was not reported in Washington and was reported only in the final 3 years of the study in California and Texas. In multivariate analyses, missing data were handled in two ways. In the first set of analyses, observations with missing responses were excluded. In the second set of analyses, missing responses were set to unknown and retained as a separate classification level. These two sets of analyses yielded similar results. Results reported in this study are for models in which observations with missing responses were excluded. The SAS statistical software package was used for all analyses.
Cause of Injury
From 1983–1991, there were 10 370 deaths attributable to injury in infants <1 year of age, resulting in an injury fatality rate of 29.72/100 000 live births (Table 1). Of these deaths, 7594 (73%) were classified as unintentional injuries, 2345 (23%) were classified as intentional injuries, and 431 (4%) were classified as undetermined intent. The leading causes of injury death in this age group, in descending order, were: homicide, suffocation, motor vehicle-related incidents, fire, drowning, food inhalation, and object inhalation. Each of these causes accounted for ≥5% the total number of injury deaths.
Over the 9-year study period, there was no significant temporal trend in the overall rate of injury death; however, there were significant trends for some specific causes of death (Fig 1). From 1983–1991, homicide rates increased by 6.4% per year (P < .001) and rates of death from mechanical suffocation increased by 3.7% per year (P < .001), whereas fatality rates from fires decreased by 4.7% per year (P < .001) and fatality rates for inhalation of objects and inhalation of food decreased by 4.8% and 4.4% per year, respectively (P < .01). There was no significant change in fatality rates for motor vehicle-related injuries or for drownings.
Associations Between Birth Characteristics and Injury Death
Characteristics of infants who died from an injury are compared with characteristics of the birth cohort and controls in Table 2. Controls were comparable with the birth cohort for all variables examined. Infants who died from an injury were more likely to be low birth weight, premature, and male. In addition, they were more likely to have mothers who were young, unmarried, with lower educational levels, higher numbers of previous live births, and late or no prenatal care. Infants born to mothers who were black or Native American were at greater risk of injury death than white infants, whereas infants born to Asian mothers were at lower risk. For all injuries combined, the highest rates were among infants born to mothers who had no prenatal care (116/100 000 live births), were <15 years of age (80/100 000), or were Native American (82/100 000).
The results of multivariate analyses for all injury deaths are shown in the last column of Table 2. Gestational age was omitted from multivariate models, because it is highly associated with birth weight and birth weight is measured more reliably. After adjusting for all other variables in the model, the following maternal characteristics increased the risk of death by twofold or more: young maternal age (OR: 2.51; 95% CI: 2.23–2.82 for 15 to 19 years vs ≥25 years and OR: 2.10; 95% CI: 1.29–3.42 for <15 years vs ≥25 years), maternal education <12 years (compared with maternal education ≥16 years; OR: 2.31; 95% CI: 2.01–2.66), no prenatal care (compared with initiation of prenatal care in the first trimester, OR: 2.11; 95% CI: 1.74–2.54), Native American race (compared with white race, OR: 2.12 95% CI: 1.63–2.75) and ≥2 previous live births (compared with no previous live births, OR: 2.14; 95% CI: 1.95–2.36).
Cause-specific injury rates, relative risks, and adjusted ORs are shown in Table 3. Key findings by the specific external cause of death are presented below.
Homicide rates were greatest in infants born to mothers who had received no prenatal care (38/100 000), were <20 years (17/100 000), single (16/100 000), black (16/100 000), or Native American (17/100 000). Homicide rates were also high among infants with birth weights <2500 g (20 and 16/100 000 in infants with birth weights <1500 g and 1500 g to 2499 g, respectively). In adjusted analyses, factors associated with an OR ≥2 included: black or Native American race, single marital status, no prenatal care, and birth weight of 1500 g to 2499 g compared with ≥2500 g.
Rates of death attributable to suffocation were also highest among infants born to mothers who had received no prenatal care (17/100 000). However, in adjusted analyses only previous live births and young maternal age were associated with a doubling of the risk of suffocation. Of the 1839 deaths in this category, 1098 (60%) were classified as accidental suffocation in a bed or cradle, 167 (9%) as suffocation by a plastic bag, and 3 (<1%) were classified as lack of air in a closed place. The remainder, 571 (31%), were in the other and unspecified categories (E-codes 913.8 and 913.9).
Motor Vehicle-Related Deaths
Infants born to mothers who were Native American were at greatest risk for motor vehicle-related deaths with a fatality rate of 14/100 000 live births. Native American race was the strongest predictor in adjusted analyses (OR: 3.6; 95% CI: 1.8–7.1).
Fires and Flames
For deaths caused by fires, the highest mortality rates were among infants born to mothers who had received no prenatal care (11/100 000), <12 years of education (8/100 000), and infants born to mothers who were black (7/100 000) or Native American (8/100 000). In adjusted analyses, each of these factors was associated with an OR ≥2, with ORs ranging from 2.1 to 2.6. However, due in part to the very low rate of death from fires in first born children, birth order was the strongest predictor of fire-related deaths. As compared with firstborn children, second and later born children had over four times the risk of dying in a fire (OR: 4.4; 95% CI: 3.2–5.9).
Rates of death from drowning were highest among infants born to mothers who had received no prenatal care (9/100 000) and among Native Americans (6/100 000). Native Americans were the highest risk group in adjusted analyses (OR: 4.8; 95% CI: 1.0–22.0). Other factors with an OR of ≥2 included; maternal age <20 years, one or more previous live births, single marital status, and no prenatal care.
Although deaths caused by inhalation of objects and food were the fifth and sixth leading cause of death attributable to unintentional injury, when viewed as a single phenomenon (choking), they combine to become the third leading cause behind mechanical suffocation and motor vehicle-related deaths. Fatality rates from choking were highest among infants with birth weights <1500 g; 15/100 000 for inhalation of objects and 10/100 000 for inhalation of food. In multivariate models, infants weighing <1500 g were at almost 12 times the risk of death caused by object inhalation, and infants weighing 1500 g to 2499 g were at three times the risk as compared with infants weighing ≥2500 g. Similar associations were seen for inhalation of food with adjusted ORs of 9.6 and 4.1 for infants weighing <1500 g and 1500 g to 2499 g, respectively. In comparison, other factors played a less important role.
During the 9-year study period 10 370 injury-related deaths were identified yielding a fatality rate of 29.72 deaths per 100 000 live births. For all injuries combined, factors associated with at least a twofold increase in the risk of death from injury included maternal age <20 years, less than a high school education, no prenatal care, Native American race, and ≥2 previous live births. The strength of associations between specific risk factors and death varied with the specific external cause of death, providing important information for targeting of cause-specific interventions.
The limitations of studies that rely on information obtained from linked birth/infant death datasets have been well documented.4 ,5 ,15 First, the only variables available for consideration are those collected on birth and death certificates. Therefore, we were unable to assess the contribution of other important factors such as household income, parenting style, the adequacy of social support networks, the use or nonuse of protective devices, or the effects of legislative and regulatory measures such as primary seatbelt laws. Second, data obtained from birth and death certificates are not always reliable. Inaccuracies are known to exist in both the assignment of the cause of death on death certificates and in the documentation of maternal and infant factors on birth certificates.16–20 Finally, the study period for the current manuscript was from 1983–1991. Linked datasets were not created for 1992–1994. The 1995–linked dataset was not included in the current analysis because of the lack of continuity that this would create and because there were methodologic differences in the manner in which the datasets were created. Results of preliminary analyses of data from the 1983– to 1986–linked datasets21 were similar to those reported in this study, confirming that associations identified in this article have remained consistent over time.
Relation to Previous Studies
In Tennessee, researchers found that low maternal education, young maternal age, and higher birth order were all associated with at least a doubling of the risk of injury-related death among children between the ages of 0 and 4 years.5 In the state of Washington, young maternal age, higher parity, black race, maternal smoking, and rural residence were independently associated with death caused by injury and, in Colorado, young maternal age and single marital status were important predictors of injury-related deaths.4 ,8 The consistency with which these sociodemographic factors are found to be associated with injuries in infants indicate that these characteristics identify high-risk populations for targeted interventions reliably. In the current study, infants born to mothers who had received no prenatal care experienced particularly high rates of death attributable to injury. Although prenatal care is unlikely to be causally linked to the risk of injury, it may serve as a proxy for a number of complex issues eg, wantedness of the pregnancy or overall use of the health care system. In any event, the first opportunity for intervention in this high-risk group may be in the hospital at the time of delivery.
Cause-Specific Analyses: Implications for Interventions
Homicide was the leading cause of death with 2345 deaths over the 9-year study (6.72/100 000 live births). This rate is probably lower than the true homicide rate as an additional 431 deaths were coded as undetermined intent, a designation that is often used for deaths that are suspicious in nature.17 In addition, a small fraction of deaths assigned to unintentional injury codes or noninjury codes may be intentional in nature.22–26 Although certain groups, such as infants born to young or single mothers were at increased risk of homicide, even in low-risk groups, homicide was generally the leading cause of traumatic death. For example, homicide was the leading cause of death for infants born to mothers who had >12 years of education or were >20 years of age. A number of primary prevention methods including home visiting, social support, and interventions aimed at improving parenting skills were reviewed recently.27 Of these, home-visiting interventions, conducted primarily in high-risk populations, showed the most promise.27–29 However, these studies focused primarily on the reduction of nonfatal child abuse and neglect, probably because fatal child abuse is relatively infrequent.30 There is a need to expand home-visiting interventions to enable evaluation of their effectiveness in preventing fatal outcomes. In addition, other approaches should be evaluated more thoroughly and new approaches should be considered.27
Native American infants experienced particularly high rates of motor vehicle-related deaths (14/100 000) and in adjusted analyses were over three times more likely to die from a motor vehicle-related death than were white infants. It is estimated that, when used properly, child safety seats reduce the risk of fatal injury in a motor vehicle crash by 69% in infants.31 Although all 50 states have laws mandating the use of child restraints, Native Americans who live on reservations are exempt from state regulations, and tribal laws vary from one reservation to another.32 In places where legislation is not sufficient, efforts should focus on the enactment of new laws or the strengthening of existing laws. To ensure that car seats are universally available, lender programs such as those used by the Navajo nation should be established (Richard Smith, Indian Health Service, written communication, May 1998). Finally, it is estimated that >80% of child safety seats are used improperly.33 Although parents and other caregivers should receive instructions on the correct use of child safety seats and the need to use the safety-seat every ride, the high rates of misuse also demonstrate the urgent need for products that are more user-friendly.
Birth order was an important predictive factor for deaths attributable to drownings, fires, and mechanical suffocation. This may be related to a decrease in parental supervision as the parents' attention is divided among more children. Alternatively, parents may overestimate the maturity and capabilities of older siblings and inappropriately expect them to supervise the infant. In two separate studies of bathtub drownings, >30% of victims were being supervised in the bathtub by an older sibling.34 ,35 Another study found that, among infants, 11% of injuries that resulted in an emergency department visit occurred while the infant was being supervised by an older sibling.36 These findings emphasize the need to counsel caregivers, not only on the risks of leaving children unattended, but also on the need for appropriate adult supervision.
Consistent with findings in the current study, researchers in Tennessee identified an independent association between maternal age, education, and birth order and fire-related deaths among children <5 years of age.10 Although the specific role of siblings in fire-related deaths is unknown, it is estimated that, in the 0- to 5-year age group, ∼40% of deaths resulting from residential fires are caused by the actions of children.37 ,38 In addition to the installation and maintenance of smoke detectors, parents should be reminded to store matches and lighters out of the reach of young children.38
One unexpected finding was the strong association between low birth weight and deaths attributable to choking. This association was seen for deaths caused by inhalation of food and for deaths caused by inhalation of objects. Of interest, 40% of deaths attributable to food inhalation and 33% of deaths caused by inhalation of objects occurred during the first 12 weeks of life, before the usual recommended age for the introduction of solid foods and before the attainment of developmental skills that would allow an infant to obtain and place an object in his/her mouth independently. Thus, although there may be a physiologic explanation for this finding, eg, oral pharyngeal coordination problems predisposing premature infants to choking episodes, it is possible that these deaths represent errors in classification. Further research is needed in this area.
From 1983–1991, the infant mortality rate in the United States decreased by 20%.11 In contrast rates of death attributable to intentional injuries and unintentional mechanical suffocation increased, and there was no change in rates of motor vehicle-related deaths or drownings. A reduction in the rate of death from injury in any one of the identified high-risk groups to the level of its comparison group would result in the saving of many lives. For example, if infants of young mothers had experienced the same rate of death from injury as infants born to mothers ≥25 years of age, 4062 lives would have been spared over the 9-year study period. Clearly, there is no single intervention that will address all causes of injury. Continued efforts should focus on the implementation of proven strategies, such as proper use of child restraints for the prevention of motor vehicle-related injuries or installation and maintenance of smoke detectors for the prevention of deaths caused by fires. For some causes of injury, such as homicide, there is a need to identify and evaluate new preventive strategies. Additionally, there is a need to identify the most effective means of communicating prevention messages to high-risk populations as well as a need to identify and reduce the barriers faced in implementing prescribed preventive strategies. Although all caregivers should receive counseling in injury prevention our findings identify high-risk subgroups for enhanced, cause-specific efforts.
- Received June 23, 1998.
- Accepted October 23, 1998.
Reprint requests to (R.A.B.) Division of Epidemiology, Statistics, and Prevention Research, National Institute of Child Health and Human Development, 6100 Executive Blvd, Room 7B03, Bethesda, MD 20892.
- ↵Anderson RN, Kochanek KD, Murphy SL. Report of Final Mortality Statistics, 1995. Monthly vital statistics report, Vol 45, No 11 (suppl 2). Hyattsville, MD: National Center for Health Statistics; 1995
- Rivara FP,
- Grossman DC
- Scholer SJ,
- Mitchel EF,
- Ray WA
- Emerick SJ,
- Foster LR,
- Campbell DT
- ↵Scholer SJ, Hickson GB, Mitchel EF, Ray WA. Predictors of mortality from fires in young children. Pediatrics 1998;101(5). URL: http://www.pediatrics.org/cgi/content/full/101/5/e12
- ↵National Center for Health Statistics. Linked Birth/Infant Death Data Set: Public Use Data File Documentation for 1983 to 1991. Hyattsville, MD; US Public Health Service: 1988–1995
- ↵World Health Organization. World Health Classification: Manual of the International Statistical Classification of Diseases, Injuries and Causes of Death, Ninth Revision. Geneva, Switzerland: World Health Organization; 1977
- ↵SAS/STAT User's Guide, Version 6, Vol 2. 4th ed. Cary, NC: SAS Institute Inc; 1990
- ↵Breslow NE, Day W. Statistical Methods in Cancer Research. Volume 1—The Analysis of Case-Control Studies. Lyon, France: IARC Scientific Publication 32. 1980
- ↵Bennett T, Braveman P, Egerter S, Kiely J. Maternal marital status as a risk factor for infant mortality. Fam Plann Perspect. 1994;26:252–256; 271
- ↵Beal SM, Byard RW. Accidental death or sudden infant death syndrome? J Pediatr Child Health. 1995:31:269–271
- Brenner R,
- Overpeck M,
- Toretsky E,
- Trumble A,
- Berendes H
- ↵Emery JL. Child abuse, sudden infant death syndrome, and unexpected infant death. Am J Dis Child. 1999;3:147:1097–1100
- Griest KJ,
- Zumwalt RE
- Kemp AM,
- Mott AM,
- Sibert JR
- ↵Harborview Injury Prevention and Research Center. Childhood Injury Prevention Interventions: Cochran Collaboration. URL: http://www.weber.u.washington.edu/hiprc/topic/abuse/cochrane.htlm
- Olds DL,
- Henderson CR,
- Chamberlin R,
- Tatelbaum R
- ↵National Highway Traffic Safety Administration. Buckle Up America—First Report to Congress. Washington, DC: US Department of Transportation, National Highway Traffic Safety Administration, Report No. DOT 888–667. 1998
- ↵Safety-belt use and motor-vehicle-related injuries, Navajo Nation, 1988–1991. Morb Mortal Wkly Rep. 1992;41:705–708
- ↵National Highway Traffic Safety Administration. Patterns of Misuse of Child Safety Seats: Final Report. Washington, DC: US Department of Transportation, National Highway Traffic Safety Administration; Report No. DOT-HS-808-440. 1996
- ↵Rauchschwalbe R, Brenner R, Smith G. The role of bathtub seats and rings in infant drowning deaths. Pediatrics 1997;100(4). URL: http://www.pediatrics.org/cgi/content/full/100/4/e1
- Rivara FP,
- Kamitsuka MD,
- Quan L
- Squires T,
- Busuttil A
- ↵Deaths resulting from residential fires—United States, 1991. MMWR Morb Mortal Wkly Rep. 1994;43:901–904
- Copyright © 1999 American Academy of Pediatrics