Although injury-related deaths have been documented among children and adult populations, insufficient attention has been directed towards injury-related infant deaths. The objective of this retrospective study was to investigate maternal and infant characteristics associated with injury-related infant deaths in West Virginia. Birth and infant mortality data for 2010–2014 were sourced from the West Virginia Bureau for Public Health, Charleston. Relative risk was calculated using log-binomial regression utilizing generalized estimating equations. Maternal characteristics associated with injury-related infant mortality in West Virginia were race/ethnicity ( = 7.48, p = .03), and smoking during pregnancy (, p < .00). Risk of a Black Non-Hispanic infant suffering an injury-related death was 4.0 (95% CL 1.7, 9.3) times that of infants of other races/ethnicities. Risk of an infant dying from an injury-related cause, if the mother smoked during pregnancy, was 2.9 (95% CL 1.6, 5.0) times the risk of such a death if maternal smoking status during pregnancy is unknown or no smoking, controlling for race/ethnicity. This study provides important information to public health stakeholders at both the state and local levels in designing interventions for partial reduction or prevention of injury-related infant mortality in West Virginia.
Citation: Koech WA, Rudisill TM, Rockett IRH (2019) Maternal characteristics associated with injury-related infant death in West Virginia, 2010-2014. PLoS ONE 14(8): e0220801. https://doi.org/10.1371/journal.pone.0220801
Editor: Marlene Camacho-Rivera, CUNY School of Medicine, City College of New York, UNITED STATES
Received: December 25, 2018; Accepted: July 23, 2019; Published: August 12, 2019
Copyright: © 2019 Koech et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying the results presented in the study can be provided to any interested party following authorization from the West Virginia Bureau for Public Health, Health Statistics Center. Please see here for the contact information for the West Virginia Bureau for Public Health, Health Statistics Center: http://www.wvdhhr.org/bph/hsc/ContactUs.asp.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Unintentional injury deaths and homicides impact children across all ages in the United States (U.S).[1–3] Although the unintentional injury-related death rate among children 19 years or less declined by 29% from 2000 to 2009 nationally, the corresponding unintentional injury-related infant mortality rate rose approximately 20%. The smallest decline in unintentional injury-related infant mortality was observed among the Black population (-21%). Between 1999 and 2004, the national unintentional injury death rate increased by 7%, and for the period 2000–2009 there was an annualized national rate of 1,081 infant deaths per 100,000 population. Overall child mortality during this period in the U.S was documented to be 11,561 children deaths per 100,000 population, whereas the rate for West Virginia was 597 per 100,000.
Overall, the rate of both unintentional and intentional injury-related deaths rose by 10% between 2000 and 2009 in the U.S. During the same period, children aged 0–4 years registered an injury-related death rate of 18.3 deaths per 100,000 population, but by 2014, unintentional injury was the fifth leading cause of infant mortality in the U.S,[8, 9] and the fourth leading cause in West Virginia.
Known risk factors associated with injury-related infant mortality include unhealthy behaviors associated with financial hardship, parental/caregiver behavior, and birth order.[11–14] A study based on California’s metropolitan areas showed an increase in the unemployment rate of 1% was associated with an increase of 8% in the unintentional injury-related infant mortality rate. Parental behaviors, such as inappropriate child car-seat placement, are associated with increased risk of unintentional injury-related infant mortality in the U.S. Similarly, caregiver behavior, such as non-adherence to infant injury prevention recommendations (e.g., not having infants sleep in the prone position or leaving an infant unattended on a raised surface), is associated with elevated risk. Head injuries are frequent among children involved in motor vehicle accidents, irrespective of age. During 2003–2013 period, unintentional injuries increased by 11%, and Black infants were two to three times more likely to die from all injury, unintentional injury and homicide compared to White infants. In the U.S., non-first born children have an increased risk of injury-related death.
Even though high rates of injury-related infant mortality are being observed in the U.S and in West Virginia, current prevention efforts in West Virginia are inadequate. The focus of this project has been maternal and infant characteristics associated with injury-related infant death in West Virginia, adjusted for race/ethnicity. Understanding maternal and infant characteristics associated with injury-related infant deaths could inform public health stakeholders on potential interventions that could partially prevent such deaths in West Virginia, an extremely poor state and the only one in the U.S. that is fully immersed in Appalachia. The Appalachian region is characterized by poor access to healthcare, low income, and high rates of unemployment and poverty.
Materials and methods
The primary data source for this analysis was birth and death records obtained from the West Virginia Bureau for Public Health, Charleston. The West Virginia University Institutional Review Board approved the parent study.
The project outcome was a binary variable labeled injury-related infant deaths (0 = alive, 1 = died). Injury-related infant deaths were constructed from external causes of mortality, building on the Centers for Disease Control and Prevention (CDC) injury classification matrix. In this matrix, injury-related infant deaths by manner of injury are: 1) accidents by cutting/piercing, drowning, fall, fire/hot object or substance, firearm, machinery, and transport; 2) homicides by cutting/piercing, drowning, fall, fire/hot object or substance, firearm, and transport; 3) other (legal intervention/war, and undetermined). Therefore, the outcome encompasses infant deaths by manner of injury, categorized by ICD-10 revision codes V01-X59, Y85-Y86, X85-Y09, Y87.01, 87.02, Y87.1, Y10-Y34, Y87.2, Y89.9, Y35, Y36, Y89.0, and Y89.1. Excluded from the outcome were infant deaths associated with other causes.
Explanatory variables included both categorical and continuous. These variables were: 1) maternal demographic, health, and behavior characteristics including education (greater than high school, high school/GED or less), insurance type/payer (Medicaid, Non-Medicaid and unknown), smoking during pregnancy (yes, no, unknown), age, race-ethnicity (Black non-Hispanic, other races/ethnicities), pregnancy term (preterm, term or post-term), trimester primary care began (first, second, third, no care), gestational diabetes (yes, no) and gestational hypertension (yes, no); and 2) newborn characteristics, including birth year, sex, and birthweight.
The statistical analysis employed both descriptive and inferential statistics. Among descriptive statistics captured were variables proportions, mean, and spread. Furthermore, statistical investigation covered variable correlation analysis using Kendell correlation and multi-collinearity was assessed.
In inferential statistics, maternal and infant characteristics associated with injury-related deaths, as well as risk of injury-related mortality, were investigated. Relative risk was calculated using log-binomial regression utilizing generalized estimating equations (GEE). Bivariable analysis was conducted, and main effects with p>0.15 were included in the model. Thereafter, a parsimonious model was developed and compared with an intercept-only model using Akaike’s Information Criterion (AIC), and two-way interaction effects for all variables were considered. In addition, model fitness was investigated using deviance. Likewise, Wald chi-square statistics and associated 95% confidence limits were reported.
Data management and analysis were conducted using SAS/STAT software, version 9.4 of the SAS systems for windows. Copyright 2002–2012 by SAS Institute Inc., Cary, NC, USA.
The West Virginia live birth distribution for the period 2010–2014 by race/ethnicity was 93.6% White (non-Hispanic), 3.5% Black (non-Hispanic), and 2.9% other races/ethnicities. A majority of injury-related infant deaths were White. When infant deaths were stratified by form of payment during the perinatal period, 41% of women utilized the Medicaid insurance program. When stratified by maternal smoking during pregnancy, 49% of the infant deaths were to the women who smoked (Table 1).
Investigation of explanatory variable correlation was implemented using the Kendell correlation, and multicollinearity was assessed using the variance inflation factor (data not shown). A bivariable binomial regression analysis was executed to identify significant main effects used in developing a fully parameterized model. Significant main effects in the bivariable regression analysis were race/ethnicity (p = .0311), smoking during pregnancy (.0006), and insurance status (p = .1198).
Significant main effects in the multivariable log-binomial regression model using GEE were race/ethnicity, = 7.48, p = .0276, and smoking during pregnancy, , p = .0003. See Table 2 for more detail. Overdispersion was assessed using model deviance, and model fitness was assessed using AIC. The intercept only model AIC was 846.1, and the parsimonious model AIC was 26.6, meaning the parsimonious model provided better predictive power than the intercept-only model.
The risk of a Black infant dying from injury-related causes was 4.0 (95% CL 1.7, 9.3) times the risk for infants from other races/ethnicities, adjusting for other factors in the model. However, the adjusted risk of an infant dying from injury-related causes if the mother smoked during pregnancy was 2.9 (95% CL 1.6, 5.0) times the risk if smoking status was unknown or the mother did not smoke (Table 2).
This study evaluated maternal and infant characteristics associated with injury-related infant death in West Virginia. Race/ethnicity and smoking during pregnancy were significantly associated with injury-related infant deaths in West Virginia. Similar findings of racial/ethnic disparity in injury-related deaths are documented by Khan and colleagues, and by Ahrens and colleagues.[3, 14] Such differences in injury-related infant mortality might be explained by racial/ethnic disparities in healthcare access and utilization. Past studies showed that Black non-Hispanics were more likely than White Non-Hispanic counterparts to receive low-quality healthcare and to have higher risk of hospital death following injury.[16, 22] Racial/ethnic disparities in injury-related infant deaths highlight the need for tailored intervention strategies that will improve healthcare access at state and local levels.
Smoking during pregnancy was significantly associated with injury-related infant mortality. This variable might be a proxy for capturing characteristics known to be associated with injury-related deaths, such as infant maltreatment. Cigarette smoking has been associated with drug abuse. Therefore, smoking during pregnancy might be a proxy for substance abuse, a risk factor for child neglect and maltreatment. It was expected to be statistically significant risk factor for infant mortality because child neglect and maltreatment is associated with drug abuse.[24, 25]
Insurance status in the perinatal period was an important maternal characteristic to investigate. Although insignificant in the multivariable analysis, it was significant in the bivariable analysis (p = 0.0566), and warrants further investigation using balanced racial/ethnic populations, since socio-economic disparities are associated with injury-related deaths by manner.
Study limitations include the small number of cases of injury-related deaths, the highly skewed nature of the racial/ethnic distribution in West Virginia, and deletion of infant deaths associated with other causes. Small numbers can destabilize statistical models. Therefore, findings from this study should be interpreted with caution. Detailed presentation of descriptive statistics, such as injury-related infant mortality rates, could not be presented. CDC opposes publication of sample size, cell sizes or mortality rates involving fewer than ten cases. The reliability of vital records might be also be problematic due to case misclassification. Past research documents sources of errors in medical records collected for purposes other than research. Finally, variables collected in vital records are limited, yielding fewer options for statistical model building. In addition, no paternal characteristics were collected for the period of pregnancy, thus the study might be missing some important albeit unknown information.
Black (non-Hispanic) infants in West Virginia have higher odds of experiencing injury-related mortality compared to infants of other races/ethnicities. Furthermore, infants of women reporting smoking during pregnancy have higher odds of injury-related death than those of women who reported not smoking or whose smoking status was unknown.
This study provides important information to public health stakeholders at both state and local levels. Study implications include formulating community-based intervention strategies to protect at-risk populations from lost years of life associated with injury-related infant mortality. Future research should investigate the association between all household characteristics and injury-related infant deaths. This study could only access maternal and infant characteristics, and paternal characteristics could provide additional important information.
Disclaimer: Findings on this article are based on research conducted by the authors and do not necessarily represent the views of West Virginia University.
The authors would like to acknowledge the West Virginia Bureau for Public Health, Charleston, West Virginia for providing the data for this study.
- 1. Fowler KA, Dahlberg LL, Haileyesus T, Gutierrez C, Bacon S. Childhood Firearm Injuries in the United States. Pediatrics. 2017;140(1).
- 2. Jain A, Khoshnood B, Lee KS, Concato J. Injury related infant death: the impact of race and birth weight. Injury prevention: journal of the International Society for Child and Adolescent Injury Prevention. 2001;7(2):135–40.
- 3. Khan SQ, Berrington de Gonzalez A, Best AF, Chen Y, Haozous EA, Rodriquez EJ, et al. Infant and Youth Mortality Trends by Race/Ethnicity and Cause of Death in the United States. JAMA pediatrics. 2018;172(12):e183317. pmid:30285034
- 4. Centers for Disease Control and Prevention. Vital signs: Unintentional injury deaths among persons aged 0–19 years—United States, 2000–2009. MMWR Morbidity and mortality weekly report. 2012;61:270–6. pmid:22513530
- 5. Centers for Disease Control and Prevention. State-specific unintentional-injury deaths—United States, 1999–2004. MMWR Morbidity and mortality weekly report. 2007;56(43):1137–40. pmid:17975527
- 6. Borse NN, Rudd RA, Dellinger AM, Sleet DA. Years of potential life lost from unintentional child and adolescent injuries—United States, 2000–2009. Journal of safety research. 2013;45:127–31. pmid:23708484
- 7. Rockett IRH, Regier MD, Kapusta ND, Coben JH, Miller TR, Hanzlick RL, et al. Leading causes of unintentional and intentional injury mortality: United States, 2000–2009. American journal of public health. 2012;102(11):e84–92. pmid:22994256
- 8. Mathews TJ, Driscoll AK. Trends in Infant Mortality in the United States, 2005–2014. NCHS data brief. 2017(279):1–8.
- 9. Kochanek KD, Murphy SL, Xu J, Tejada-Vera B. Deaths: Final Data for 2014. National vital statistics reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics, National Vital Statistics System. 2016;65(4):1–122.
- 10. West Virginia Bureau for Public Health. West Virginia vital statistics, 2014. In: Resources WVDoHaH, editor. Charleston WV: West Virginia Department of Health and Human Resources, Bureau for Public Health, Health Statistics Center; 2016. p. 184.
- 11. Arbogast KB, Wozniak S, Locey CM, Maltese MR, Zonfrillo MR. Head impact contact points for restrained child occupants. Traffic injury prevention. 2012;13(2):172–81. pmid:22458796
- 12. Bruckner TA. Metropolitan economic decline and infant mortality due to unintentional injury. Accident; analysis and prevention. 2008;40(6):1797–803. pmid:19068279
- 13. Heerman WJ, Perrin EM, Sanders LM, Yin HS, Coyne-Beasley T, Bronaugh AB, et al. Racial and Ethnic Differences in Injury Prevention Behaviors Among Caregivers of Infants. American journal of preventive medicine. 2016;51(4):411–8. pmid:27291075
- 14. Ahrens KA, Rossen LM, Thoma ME, Warner M, Simon AE. Birth Order and Injury-Related Infant Mortality in the U.S. American journal of preventive medicine. 2017;53(4):412–20. pmid:28666774
- 15. Mathews TJ, MacDorman MF, Thoma ME. Infant mortality statistics from the 2013 period linked birth/infant death data set. National vital statistics reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics, National Vital Statistics System. 2015;64(9):1–30.
- 16. Bernard SJ, Paulozzi LJ, Wallace DL, Centers for Disease C, Prevention. Fatal injuries among children by race and ethnicity—United States, 1999–2002. Morbidity and mortality weekly report Surveillance summaries. 2007;56(5):1–16. pmid:17510613
- 17. Black DA, Labor market performance, poverty and income inequality in appalachia. Appalachian Regional Commission 2004.
- 18. Annest J, Hedegaard H, Chen L, Warner M, Small E. Proposed Framework for Presenting Injury Data using ICD-10-CM External Cause of Injury Codes. 2014:1–32.
- 19. McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. American journal of epidemiology. 2003;157(10):940–3. pmid:12746247
- 20. SAS Institute. SAS/STAT software SAS Institute Inc; 2012.
- 21. Rao CR, Wu Y. On model selection: IMS Lecture Notes—Monograph Series 2001;38.
- 22. Arthur M, Hedges JR, Newgard CD, Diggs BS, Mullins RJ. Racial disparities in mortality among adults hospitalized after injury. Medical care. 2008;46(2):192–9. pmid:18219248
- 23. King AJ, Farst KJ, Jaeger MW, Onukwube JI, Robbins JM. Maltreatment-Related Emergency Department Visits Among Children 0 to 3 Years Old in the United States. Child maltreatment. 2015;20(3):151–61. pmid:25631298
- 24. Lai S, Lai H, Page JB, McCoy CB. The association between cigarette smoking and drug abuse in the United States. Journal of addictive diseases. 2000;19(4):11–24. pmid:11110061
- 25. Smith DK, Johnson AB, Pears KC, Fisher PA, DeGarmo DS. Child maltreatment and foster care: unpacking the effects of prenatal and postnatal parental substance use. Child maltreatment. 2007;12(2):150–60. pmid:17446568
- 26. Cubbin C, Smith GS. Socioeconomic inequalities in injury: critical issues in design and analysis. Annual review of public health. 2002;23:349–75. pmid:11910067
- 27. The United States Cancer Statistics. National Center for Health Statistics staff manual on confidentiality: Centers for Disease Control and Prevention; 2004 . Available from: https://www.cdc.gov/cancer/uscs/technical_notes/stat_methods/suppression.htm.
- 28. Northam S, Knapp TR. The reliability and validity of birth certificates. J Obstet Gynecol Neonatal Nurs. 2006;35(1):3–12. pmid:16466348