Pregnancy-associated homicide (homicide occurring during or within 1 year of pregnancy) has been estimated as ranging from 8.4% to 43.3% of all pregnancy-associated mortality.1–6 The Centers for Disease Control and Prevention reported an overall pregnancy-associated homicide rate of 1.7 per 100,000 live births for 1991–1999.6 Among 16 states reporting to the National Violent Death Reporting System from 2003 to 2007, the pregnancy-associated homicide rate was 2.9 per 100,000 live births, a higher rate than for specific direct obstetric causes (hemorrhage, hypertensive disorders, or amniotic fluid embolism).7
Scant research has addressed whether women who are pregnant or within 1 year of pregnancy are at increased risk for homicide. In New York and Washington, DC, in the 1980s and 1990s, no association was observed between pregnancy status and homicide risk.1,5 In North Carolina, pregnancy and postpartum status were strongly protective against homicide among women 30 years and older (rate ratio 0.17 [95% confidence interval (CI) 0.04–0.68]).
In Illinois between 1986 and 1989, Cook County medical examiner records showed that 25 of 95 maternal deaths (26.3%) were the result of homicide, more than any other single cause.8 One Illinois perinatal network in the 1990s found that homicide caused 4 of 42 maternal deaths (9.5%).9 The purpose of this study was to examine whether being pregnant or within a year of pregnancy was associated with excess risk for homicide compared with other females of reproductive age in Illinois between 2002 and 2011 and to describe the association by race, ethnicity, and age group.
MATERIALS AND METHODS
This is a retrospective, multicohort, ecologic analysis of maternal deaths resulting from homicide between 2002 and 2011 in Illinois. In Illinois, maternal death is defined as “the death of any woman dying of any cause whatsoever while pregnant or within 1 year after the termination of the pregnancy, irrespective of the duration of the pregnancy at the time of the termination or the method by which it was terminated.”10 The Illinois Department of Public Health identifies maternal deaths in several ways, including direct notification by the hospital where the death occurred, a death certificate checkbox indicating that the decedent had been pregnant within 1 year of death, vital records searches, newspaper articles, and obituaries. All known maternal deaths are reviewed by 1 of the 10 regional perinatal centers that oversee maternity care at all hospitals in the state. The process of perinatal center review has been described in greater detail elsewhere.11,12 The institutional review boards at the Illinois Department of Public Health and the University of Illinois at Chicago determined that this research did not involve “human subjects” as defined in 45 CFR 46.102(f).
This study uses data from three sources. The data for maternal deaths are drawn from the Illinois Department of Public Health maternal mortality review database that contains data from the death certificate, associated birth certificate or fetal death certificate, medical records, and the results of perinatal center review. The data are entered by the Illinois Department of Public Health staff member who is responsible for maternal mortality data, the Perinatal Quality Control Inspector. The Perinatal Quality Control Inspector enters data on decedent characteristics based on collected records before sending the case to the perinatal center for review. After the review, all records are returned to the Perinatal Quality Control Inspector and the results of the review are entered into the database. Before designating a case as completed, the Perinatal Quality Control Inspector ensures that data in the database match all data on the form. If a form contains conflicting or implausible data, the Perinatal Quality Control Inspector returns the form to the perinatal center for correction before entering the data in the database.
Data on the number of homicides of females of childbearing age (10–49 years) in Illinois by race and ethnicity for the same time period are drawn from the Centers for Disease Control and Prevention Web-based Injury Statistics Query and Reporting System, an interactive online database that provides data associated with fatal and nonfatal injury and violent death.13 Data for the total population of females of childbearing age (10–49 years) were also obtained from Web-based Injury Statistics Query and Reporting System. The total number of live births during the study period classified by race, ethnicity, and age group was provided to us by the Illinois Department of Public Health (private communication).
At the time of this analysis, there were 708 cases of maternal death in the maternal mortality review database from 2002 to 2014. As a result of the length of time required for collection of records and completion of perinatal center review, data from 2012 through 2014 were incomplete and were excluded from analysis, resulting in a final analysis sample of 636. The maternal mortality review database includes demographics, determination of the cause of death as related to pregnancy, an assessment of the potential preventability of the death, and identification of avoidable factors (patient, health care provider, or systems).11
We estimated the population of nonpregnant and nonpostpartum females in the state and the number of homicides among these females using data from the maternal mortality review database, Web-based Injury Statistics Query and Reporting System, and vital records. We subtracted the number of known pregnancy-associated homicides recorded in the maternal mortality review database from the total number of homicides among Illinois females aged 10–49 years that was recorded in the Web-based Injury Statistics Query and Reporting System to identify homicides that were not to pregnant or postpartum females. Likewise, we subtracted the number of live births provided to us by the Illinois Department of Public Health from the total population of females ages 10–49 years recorded in the Web-based Injury Statistics Query and Reporting System to calculate the total number of females who were not pregnant or postpartum.
We used χ2 tests to compare the distributions of age, race, ethnicity, and county of residence for victims of pregnancy-associated homicides and all females with live births in the state. We calculated mortality rates per 100,000 live births for pregnancy-associated homicide, other causes of pregnancy-associated death related to violence and injury, and selected pregnancy-related causes of death. We examined observed 10-year homicide rates by pregnancy status overall and stratified by both age group and race and ethnicity. We assessed stratum-specific relative risks using χ2 tests. In addition, Mantel-Haenszel tests for adjusted measures and Breslow-Day tests for homogeneity were used as appropriate. We used SAS 9.3 for all data analysis.
In Illinois, of the 636 known deaths to females while pregnant or within 1 year of pregnancy, there were 82 (12.9%) pregnancy-associated homicides. Approximately 38% of the homicides occurred during pregnancy, 18% occurred in the first 6 weeks postpartum, and 44% occurred in the late postpartum period (from 43 to 365 days after delivery; Table 1). More than half of the homicides were of women aged 20–29 years (n=53 [64.6%]), non-Hispanic black women (n=43 [52.4%]), women residing in Cook County (n=47 [57.3%]), and unmarried women (n=57 [69.5%]). The perinatal center conducting the review determined that 62.2% (n=51) of the homicides were potentially preventable. Potentially preventable patient factors (eg, late entry to prenatal care, alcohol use, and smoking) were cited in nine homicides (11.0%) and systems factors (eg, domestic and community violence) were cited in seven homicides (8.5%); no health care provider factors (eg, delay in diagnosis or treatment) were identified for any pregnancy-associated homicides (data not shown).
Figure 1 shows the pregnancy-associated mortality rates from both clinical and nonclinical causes. Over the study period, the pregnancy-associated homicide rate is statistically significantly higher than the mortality rate from any of the four leading causes of death directly related to pregnancy (hemorrhage, emboli, severe preeclampsia and eclampsia, and sepsis). Of the 636 pregnancy-associated deaths during the study period, 95 (14.9%) were caused by motor vehicle crashes, 40 (6.3%) resulted from suicide, and 36 (5.7%) resulted from substance abuse. Death from motor vehicle crashes is the only single cause of pregnancy-associated mortality more common than homicide.
The overall pregnancy-associated homicide rate during the study period was 5.03 per 100,000 live births (95% CI 4.02–6.22), whereas the homicide rate among females who were not pregnant or within a year of pregnancy was 2.88 per 100,000 females (95% CI 2.70–3.07; Table 2). After stratifying by age group, the overall pregnancy-associated homicide rate was 6.53 per 100,000 live births among females ages 10–29 years (95% CI 5.09–8.26) and 2.14 per 100,000 live births among those ages 30–49 years (95% CI 1.24–3.45).
For each stratum of pregnancy status and race or ethnicity, the highest homicide rates were observed among females ages 10–29 years who were pregnant or within 1 year of pregnancy. The absolute increases in homicide associated with pregnancy among females ages 10–29 years were 1.62, 9.49, and 2.56 homicides per 100,000 females for non-Hispanic white, non-Hispanic black, and Hispanic females, respectively. Regardless of pregnancy status, homicide rates among Illinois females are highest among non-Hispanic black females (8.56/100,000 [95% CI 7.58–9.64] for ages 10–29 years and 9.41/100,000 [95% CI 8.32–10.60] for ages 30–49 years).
The relative risk for the association between homicide and pregnancy adjusted by race or ethnicity was 2.20 (95% CI 1.70–2.85) and 0.88 (95% CI 0.53–1.47) for females ages 10–29 and 30–49 years, respectively (Table 3). The more than twofold increase in risk among younger females did not differ by race or ethnicity (Breslow-Day χ2 0.0085, P=.99). Among both non-Hispanic blacks and whites, however, the association between pregnancy and homicide differed by age: there is approximately a twofold increase in risk for younger females (10–29 years) compared with their nonpregnant counterparts, whereas the observed rates suggest a decreased risk of homicide for older pregnant women 30–49 years compared with nonpregnant and nonpostpartum women (Breslow-Day χ2 7.01, P<.01, and 5.14, P=.02, respectively). In contrast, among Hispanic females, there is approximately a twofold increase in risk of homicide for pregnant or recently pregnant females in both age groups compared with nonpregnant and nonpostpartum females (Breslow-Day χ2 0.01, P=.92).
We found a pregnancy-associated homicide rate in Illinois from 2002 to 2011 of 5.0 per 100,000 live births, accounting for approximately 13% of pregnancy-associated deaths in the state. Being pregnant or within 1 year of pregnancy more than doubled the risk of homicide among females younger than 30 years of age compared with nonpregnant or recently pregnant females of the same age. Non-Hispanic black and Hispanic females were at higher risk of homicide than non-Hispanic white females regardless of age group or pregnancy status.
We observed relative risks for the association between pregnancy and homicide suggestive of a protective effect of pregnancy among non-Hispanic white and black women ages 30–49 years, but pregnancy remained a risk factor for homicide among older Hispanic women. Although these estimates did not reach statistical significance, this is an interesting finding for which we do not have a ready explanation. The complex interactions across age, race, and ethnicity should be examined more fully, perhaps using multistate data to achieve greater statistical power.
We found both a higher pregnancy-associated homicide rate (5.0/100,000 live births compared with 2.9) and a greater proportion of pregnancy-associated homicides in the postpartum period (56% compared with 22%) compared with Palladino's assessment of 16 states from 2003 to 2007 using the National Violent Death Reporting System.7 We believe these discrepancies are the result of inaccurate ascertainment of female homicide victims in the first year postpartum attributable to the large proportion (67.2%) of female deaths in the National Violent Death Reporting System coded as “unknown pregnancy or postpartum status.”7
Our study has several limitations. First, we were unable to definitively ascertain the exact number of females in Illinois who were pregnant or recently pregnant because we may have missed those who experienced abortion (spontaneous or induced), a limitation inherent to a multicohort study. Although the inability to include every pregnant or postpartum female in the denominator may have overestimated the association between pregnancy and homicide, pregnancy-associated homicides may also be misclassified as having occurred to females who had not experienced pregnancy within a year, particularly among females in the late postpartum period, resulting in an underestimate. Additionally, we were unable to determine the means of homicide death or whether the homicides were related to intimate partner violence, other interpersonal conflicts, or random events. Finally, without a richer source of contextual information for all cases, it is impossible to determine what role, if any, pregnancy itself played in the circumstances leading to these homicides.
Notwithstanding these limitations, our study contributes yet another example to a growing body of literature about health disparities for black and Hispanic females compared with white females. Pregnancy-associated homicide in Illinois follows the same demographic trends by race or ethnicity and age as seen in the U.S. population overall, namely that homicide rates are highest among non-Hispanic black and Hispanic people, younger people, and people residing in urban areas.13–16 This speaks to the multifaceted nature of violence in our society.
In Illinois, the statewide maternal mortality review committee focuses on deaths resulting from obstetric and other clinical causes; however, we found mortality rates for maternal death resulting from violence and injury higher than rates for the most common obstetric causes. Homicide, motor vehicle crashes, suicide, and substance abuse or overdose comprised more than one third of deaths of females while pregnant or within a year of pregnancy in Illinois during the study period. To address this issue, the Illinois Department of Public Health created a second statewide maternal mortality review committee to conduct multidisciplinary reviews of pregnancy-associated violent deaths. The new committee has the potential to remedy our last two limitations by using different sources of information and a data collection form tailored to capture information relevant to violent death. The committee reviews these cases in depth and identifies potentially modifiable factors and interventions that could prevent these deaths.
Intimate partner violence is known to be a major factor associated with female homicide deaths in the United States15,16 as well as among pregnancy-associated homicide deaths.7,17 Our findings underscore the need for obstetric health care providers and community agencies to offer quality screening and treatment services to females during this interval.
The American College of Obstetricians and Gynecologists recommends screening all obstetric patients for intimate partner violence at the first prenatal visit, at least once per trimester, and at the postpartum checkup.18 However, there are barriers to health care provider uptake of screening for intimate partner violence, such as lack of time and lack of availability of quality services for females who do screen positive.19–22 In Illinois, screening for perinatal depression is mandated by law and screening is billable as a “risk assessment.” A similar mandate could allow health care providers to be reimbursed for intimate partner violence screening. Community awareness and mobilization are required for this to change at a systemic level.
More than half of pregnancy-associated homicides in Illinois were in women aged 20–29 years, non-Hispanic black women, women residing in Cook County, and unmarried women. We found evidence that pregnancy and the postpartum period are associated with additional risk of homicide among females younger than 30 years. Although all violence against females must be addressed, we recommend that state maternal mortality review committees, in addition to reviewing deaths resulting from obstetric and clinical causes, should conduct in-depth reviews of pregnancy-associated homicides and other violent deaths. It is only when we assess factors that increase females’ risk of homicide during pregnancy and the postpartum period that we can identify populations and targets amenable to intervention.
1. Krulewitch CJ, Pierre-louis ML, de Leon-Gomez R, Guy R, Green R. Hidden from view: violent deaths among pregnant women in the District of Columbia, 1988-1996. J Midwifery Womens Health 2001;46:4–10.
2. Horon IL, Cheng D. Enhanced surveillance for pregnancy-associated mortality—Maryland, 1993-1998. JAMA 2001;285:1455–9.
3. Dietz PM, Rochat RW, Thompson BL, Berg CJ, Griffin GW. Differences in the risk of homicide and other fatal injuries between postpartum women and other women of childbearing age: implications for prevention. Am J Public Health 1998;88:641–3.
4. Jocums SB, Berg CJ, Entman SS, Mitchell EF Jr. Postdelivery mortality in Tennessee, 1989-1991. Obstet Gynecol 1998;91:766–70.
5. Dannenberg AL, Carter DM, Lawson HW, Ashton DM, Dorfman SF, Graham EH. Homicide and other injuries as causes of maternal death in New York City, 1987 through 1991. Am J Obstet Gynecol 1995;172:1557–64.
6. Chang J, Berg CJ, Saltzman LE, Herndon J. Homicide: a leading cause of injury deaths among pregnant and postpartum women in the United States, 1991-1999. Am J Public Health 2005;95:471–7.
7. Palladino CL, Singh V, Campbell J, Flynn H, Gold KJ. Homicide and suicide during the perinatal period: findings from the National Violent Death Reporting System. Obstet Gynecol 2011;118:1056–63.
8. Fildes J, Reed L, Jones N, Martin M, Barrett J. Trauma: the leading cause of maternal death. J Trauma 1992;32:643–5.
9. Panting-Kemp A, Geller SE, Nguyen T, Simonson L, Nuwayhid B, Castro L. Maternal deaths in an urban perinatal network, 1992-1998. Am J Obstet Gynecol 2000;183:1207–12.
10. Illinois General Assembly. Section 657.10: definition of maternal death. 25 Ill Reg 16491, effective January 1, 2002.
11. Geller SE, Koch AR, Martin NJ, Rosenberg D, Bigger HR; Illinois Department of Public Health Maternal Mortality Review Committee Working Group. Assessing preventability of maternal mortality in Illinois: 2002-2012. Am J Obstet Gynecol 2014;211:698.e1–11.
12. Geller SE, Koch AR, Martin NJ, Rosenberg D; Illinois Department of Public Health Maternal Morbidity Review Committee Working Group. Comparing two review processes for determination of preventability of maternal mortality in Illinois. Matern Child Health J 2015;19:2621–6.
13. Centers for Disease Control and Prevention. Web-based injury statistics query and reporting system. 2014. Available at: www.cdc.gov/injury/wisqars/
. Retrieved January 15, 2015.
15. Smith EL, Cooper A. Homicide in the U.S. known to law enforcement, 2011. Washington, DC: U.S. Department of Justice. 2013.
17. Martin SL, Macy RJ, Sullivan K, Magee ML. Pregnancy-associated violent deaths: the role of intimate partner violence. Trauma Violence Abuse 2007;8:135–48.
18. Intimate partner violence. Committee Opinion No. 518. American College of Obstetricians and Gynecologists. Obstet Gynecol 2012;119:412–7.
19. Sprague S, Madden K, Simunovic N, Godin K, Pham N, Bhandari M, et al. Barriers to screening for intimate partner violence. Women Health 2012;52:587–605.
20. Calderón SH, Gilbert P, Jackson R, Kohn MA, Gerbert B. Cueing prenatal providers. Effects on discussions of intimate partner violence. Am J Prev Med 2008;34:134–7.
21. Humphreys J, Tsoh JY, Kohn MA, Gerbert B. Increasing discussions of intimate partner violence in prenatal care using Video Doctor plus Provider Cueing: a randomized, controlled trial. Womens Health Issues 2011;21:136–44.
22. O'Campo P, Kirst M, Tsamis C, Chambers C, Ahmad F. Implementing successful intimate partner violence screening programs in health care settings: evidence generated from a realist-informed systematic review. Soc Sci Med 2011;72:855–66.