|Home | About | Journals | Submit | Contact Us | Français|
Although there has been much research examining the relationship between pregnancy and abuse, this study is one of the few to investigate whether perinatal status (defined as pregnancy or early postpartum) impacts the help seeking of abused women.
We retrospectively reviewed 3 years of prosecutor administrative records, police incident reports, and hospital medical records for a countywide population of adult females (n=964) assaulted by an intimate partner in 2000. Perinatal and nonperinatal victims were compared using chi-square and a series of logistic regression models, controlling for all demographic and incident-related factors.
Compared with women across the county, abused women were twice as likely to become pregnant (p<0.001). Perinatal status did not change the rate of help seeking from police (OR 1.1, p=0.67) or emergency departments (ED) (OR 1.1, p=0.94), but it did change the pattern of help seeking with higher ED use in the 6 months prior to the assault (p<0.01) and a trend toward seeking help with fewer injuries (p=0.10).
Abused women are more likely to become pregnant. Perinatal status impacts how victims seek help from criminal justice agencies and EDs.
Intimate partner violence (IPV) has considerable adverse effects on maternal and fetal outcomes.1 As the leading nonobstetrical cause of maternal injury and death,2,3 IPV is strongly associated with fetal-infant morbidity and mortality.4–8 In addition to direct physical trauma, IPV has been shown to inflict harm indirectly through its impact on prenatal healthcare, maternal substance use, and abuse-related mental health sequelae.9–12
Women in the perinatal period, defined as pregnancy and early postpartum, are considered to be at higher risk of physical violence by an intimate partner. Gazmararian et al.13 found that between 3.9% and 8.3% of pregnant women experience IPV, a rate at least three times higher than the generally reported 1.3% physical and sexual violence found among adult women in population-based surveys.14 Historically, this has been attributed to an increase in violence during pregnancy.3,15 More recently, however, researchers have suggested that this greater rate of victimization among pregnant women may be due, at least in part, to a higher pregnancy rate among abused women.11,16 Understanding the source of this higher prevalence has significant implications for public policies and programs serving women's reproductive health.
Although there has been much research studying the relationship between pregnancy and abuse, there has been less research investigating the relationship of pregnancy to help seeking by IPV victims. Existing literature about help seeking by IPV victims in general has ranged from learned helplessness, where victims are viewed as debilitated from the abuse and unable to help themselves,17 to evidence that women actively use services and strategies to both cope with and escape from the violence in their lives,18,19 including seeking support from friends and family, calling the police, seeking medical care, applying for protection orders, and leaving the abuser.20 Many studies illustrate the complexity of help seeking, citing the influence of relationship status (whether currently together or separated), the emotional and practical resources available, and the nature of the abuse (type, severity, length of abuse) as factors mitigating victims' help seeking.21–24 Perinatal status may further complicate victim outreach, although there is little research on how perinatal women's help seeking may vary from that of nonperinatal women.
This study explores the relationship between perinatal status and IPV in a high-risk population of women who have been victims of IPV resulting in a police request for charges. We compare characteristics of perinatal women with nonperinatal women, focusing on their help-seeking behavior. Specifically, we examine the relative use of police and emergency departments (EDs) after an IPV assault. Understanding whether perinatal women seek help differently, with different rates or different patterns of use, is critical to guiding outreach and service efforts to this vulnerable population.
This study examined all criminal justice and hospital records for an entire, countywide population of adult, female IPV assault victims throughout the study index year (2000). This involved a retrospective review of administrative records from the county prosecutor's office, police incident reports, and hospital medical records. Records from 1999 through 2001 were reviewed. The prosecutor dataset provided demographic and case information about the IPV assault. The police dataset identified whether or not the victim initiated the police call for help, and the hospital dataset provided information about the victim's perinatal status at the time of the assault, as well as whether she sought medical help from the ED (and characteristics of that visit).
The study was conducted in a county located in southwest Michigan, with a total population of 238,603 in the year 2000; 98,192 of the population were adult women, aged ≥16.25 Of these, 964 (0.98%) were victims in IPV assault cases that rose to the level of a crime in the year 2000. Our study sample, then, was this cohort of adult women, aged ≥16, who were victims in at least one assault charge from their intimate partner (defined as spouse or former spouse, current or past dating partner, or had a child in common).
The study population was victims involved in a total of 1438 IPV assaults leading to charging requests over the 3-year study period. These 1438 incidents formed the basis of our analysis. The prosecutor database contained victim identifiers (full name, date of birth, address) as well as incident descriptors (date charge requested, approval for prosecution, misdemeanor or felony, case disposition, injury noted by law enforcement, and victim-defendant relationship). For each of the 1438 incidents, the corresponding police incident report was reviewed to determine whether or not the victim had been the one to call the police for help. Hospital electronic medical records were searched to identify perinatal status and department visits associated with the assault (date, chief complaint, whether ED staff identified IPV).
Police help seeking was defined by whether or not women called police themselves. Additionally, women who directed others (often their children) to call police were also coded as help seeking from police. ED help seeking was coded Yes if the visit occurred within 1 day of the known police-identified assault, even if ED staff were unaware that the visit was IPV related. Perinatal status was defined as being either pregnant or postpartum at the time of a charge-requested incident. Because this study considered victim decision making, the pregnancy variable was predicated on an assumption of pregnancy knowledge. Thus, the pregnancy period was defined starting at 4 weeks gestation rather than at the beginning of gestation and ending at pregnancy termination (either delivery or miscarriage/abortion). The postpartum period was defined from the time of termination (again, either delivery or miscarriage/abortion) until 12 weeks postpartum. Of the 964 women, 83 were found to be pregnant at the time of an assault, and 36 were postpartum.
Finally, U.S. Census data were used to generate national and county comparisons for demographic and pregnancy rate information.25 For a fuller description of data collection methods, refer to Kothari and Rhodes.26
Access and permission to use the prosecutor administrative database were granted by the Kalamazoo County Prosecuting Attorney's Office and facilitated by the Michigan Department of Community Health. As a public health surveillance study, a waiver of consent for medical record review and a Health Insurance Portability and Accountability Act established standards protecting the privacy of health information, (HIPAA) waiver were approved by the Institutional Review Boards (IRB). In order to preserve confidentiality, all identifiers were removed from the datasets once data collection was completed.
Our data analysis began by comparing demographics of our study population to that of the United States and county on age, race, and marital status. The statistical significance of our study population compared with the county from which it was drawn was tested with chi-square. Then, within our study population, we compared the characteristics of women who were pregnant or postpartum (perinatal) at the time of assault vs. those who were not. This comparison was also tested using chi-square. Finally, we calculated the proportions of the perinatal and nonperinatal groups who called the police and visited the ED. These rates were compared overall and were stratified by demographic and incident-related factors. We also produced a series of logistic regression models to examine the relationship between perinatal status and help seeking (from police and from EDs), controlling for all demographic and incident-related factors. We also ran logistic models that included an interaction between perinatal status and each demographic and incident factor to determine if the odds ratio (OR) differed in magnitude across strata. All statistical analyses were conducted using Proc Surveyfreq and Surveylogistic in SAS v. 9.13 (SAS Institute, Cary, NC) to account for the nesting of multiple incidents within subjects.
Table 1 presents demographic information for the study population. Compared with adult women across the nation, women in this county tend to be younger, Caucasian, and single. Within the county, the study population of women victims in IPV charging requests was more likely to be in the prime childbearing years (aged 20–34) than the county population and was disproportionately African American and single.
As noted previously, the 964 study women were victims in 1438 IPV assault-related charging requests; 68.6% in a single incident and the remaining 31.4% in multiple incidents, from two to eight, over the 3-year study period. These are cases where a law enforcement agency has responded to an assault call and subsequently submitted a charging request to the prosecutor's office. Of the 1438 incidents, nearly 80% (n=1108) of charging requests resulted in approval, meaning either the suspected abuser was charged with a crime and arrested or a warrant for arrest was issued. Over two thirds (n=768) of these approved cases resulted in conviction, through either plea bargaining or trial. Nearly 60% of the total incidents (n=815) involved some form of physical injury, as noted by the responding officer. Of these, 95% (n=774) were minor injuries (e.g., cuts and bruises), 4.9% (n=40) were major injuries (e.g., loss of consciousness, internal injury, broken bones), and 1 incident was fatal.
The rate of pregnancy among the study population was nearly double that of the county population of women of childbearing age (18.0% of study population pregnant vs. 8.7% of county population pregnant in the year 2000).25 Despite this high rate of pregnancy during the year 2000, only 7.8% were pregnant at the time of the charged assault. A little over 3% (3.2%) were assaulted during early postpartum.
Table 2 shows that compared with their counterparts, study women who were pregnant or postpartum (perinatal) at the time of the assault were significantly different from nonperinatal women on several measures. They were younger (25.9% were teenagers), more likely to be African American, and more likely to be in a dating relationship with their abuser rather than married to him. Furthermore, dating perinatal women were also more likely to be separated from their boyfriend at the time of the assault compared with nonperinatal women.
As shown in Table 2, perinatal women and nonperinatal women were equally likely to disclose a previous history of assault to court-based victim advocates. In addition, they were just as likely to have a history of criminal justice intervention for these assaults. However, perinatal women were significantly more likely than nonperinatal women to have recently (within the last 6 months) visited an ED.
The results shown in Table 2 illustrate a statistical trend toward perinatal women having fewer observable injuries than nonperinatal women. Type of injury—either major, such as broken bones, or minor, such as cuts or bruises—was not significantly different between the two groups. Case dispositions varied considerably between perinatal and nonperinatal women. Perinatal women were more likely to have charges approved by the prosecutor initially but also to have those charges later dropped, by either the prosecutor or the court. In the end, there was a net loss, with fewer cases involving perinatal women resulting in conviction.
The results of the multivariate analysis (Table 3) demonstrate that perinatal status had little overall effect on help seeking. Women who were pregnant or postpartum when assaulted were just as likely to call police or visit an ED as women who were not pregnant or postpartum. Both measures, though, have wide confidence intervals (CI), indicating great variability within each perinatal group.
There were, however, suggestive interaction effects for victim demographics (age and race) and calling police. Table 3 shows that perinatal women who were older (age 34+) trend toward being less likely to call police than nonperinatal women who were older (age 34+). Similarly, perinatal women who were Caucasian trend toward less police help seeking than nonperinatal women who were Caucasian. Furthermore, perinatal women with more severe incidents (multiple charges and felony level) were twice as likely (OR 2.2) to call police compared with nonperinatal women.
None of the interaction effects for victim/incident factors and ED help seeking reached significance, being characterized by low overall rates of ED help seeking and great variability among the small set with assault-related ED visits. Given this, there may be a tendency for perinatal groups that are less likely to call police to be more likely to visit an ED. This was evident for perinatal women who were older, Caucasian, and privately insured, whereas perinatal women using the criminal justice system, typically those with more severe assaults (with multiple charges and felony-level charges), may have a tendency to use the ED less (OR 2.2 for perinatal women calling police for multiple, felony-level charges compared with OR 0.2 for perinatal women visiting an ED for these same incidents).
Although the sample was too small for statistical comparison (n=15 perinatal ED help seekers, n=119 nonperinatal ED help seekers), there were notable differences between the two groups regarding their ED presentation. Both cited assault as the most common reason for coming to the ED (6 of 13[46%] perinatal and 52 of 119[44%] nonperinatal), but perinatal women were more likely to give a reproductive-related complaint (abdominal pain, pregnancy concerns, bleeding) (3 of 13[23%] perinatal women vs. 7 of 119[6%] nonperinatal). Perinatal women were identified by ED staff as IPV victims less frequently than nonperinatal women, with only 54% (7 of 13) visits with IPV documentation compared with 71% (85 of 119) visits of nonperinatal women.
Our study adds to the existing literature with the following key findings: (1) abused women are more likely to become pregnant, and (2) being pregnant or postpartum influences their pattern of help seeking from criminal justice and ED systems but not their overall rate. Compared with nonperinatal women, perinatal women seek help with fewer observable injuries, are more likely to have visited the ED in the last 6 months, and have higher charging authorization but lower final conviction rates.
This study demonstrates yet again the vulnerability of IPV victims; not only were they younger, members of a minority, and single, but also they were twice as likely to be pregnant compared with their age-matched neighbors. Although some have speculated that pregnancy may be the precipitating factor for IPV, this finding suggests the reverse is more likely: that IPV predisposes a woman to pregnancy. Additionally, perinatal victims in our study reported violence prior to the current assault just as often as nonperinatal victims. Concluding that IPV leads to more pregnancy is consistent with other evidence that abused women are more likely to have unintended pregnancies27,28 as a result of living in an abusive relationship, with less control over use of condoms or other birth control.29
This study further documents that perinatal victims interact differently from nonperinatal victims with criminal justice and EDs. Despite fewer injuries, perinatal victims sought help from police and EDs at rates similar to that of nonperinatal victims. IPV help seeking generally is correlated with increasing severity20,30 rather than reduced severity, and this could mean that perinatal women have a lower threshold for seeking help. Also, our study provides evidence that once in the criminal justice system, such women may be more motivated participants, at least initially. Perinatal study victims were more likely to have charges authorized than were nonperinatal victims. Because charging authorization, especially with no documented injury, is often dependent on a victim providing evidence for the prosecution,31 perinatal women maybe be more actively engaged at this point in adjudication than nonperinatal women. However, that fewer perinatal cases made it to trial and even fewer to conviction suggests that perinatal victim participation waned over the course of adjudication. Studies by qualitative researchers have documented that women find it much harder to leave an abusive partner during pregnancy and the postpartum period for several reasons: increased social pressure to stay together, fear of being the sole parent and financial provider, and a natural inclination to reinvest in the relationship for the sake of the infant.32,33
Not only are perinatal women as likely as nonperinatal women to seek help from EDs for the known assaults, but they are also more likely to have used EDs in the previous 6 months. This supports other research findings that abused pregnant women have higher healthcare utilization than nonabused pregnant women.34 Clearly, many opportunities for identification and referral of this vulnerable group exist within the healthcare setting, but similar to primary care settings, few of our perinatal study victims actually disclosed their abuse to ED staff. Women, in general, rarely disclose abuse to healthcare providers unless actively and sensitively screened,35,36 an occurrence that is unfortunately rare in busy EDs.37 This presents a missed opportunity because women are open to discussing abuse with providers who express compassion and who can help link them to community resources.35 Certainly, among nonpregnant abused women, this offers a chance to empower women regarding their reproductive rights and health.
Finally, findings from our study indicate that perinatal status has a differential impact on help seeking depending on the age and race of a woman, with older and Caucasian perinatal women less likely to call police than their nonperinatal counterparts and possibly more likely to seek help from EDs. Other studies have documented the impact of demographics, especially age, on the prevalence of abuse during pregnancy,38 but this study also shows the impact of age on a pregnant woman's response to that abuse.
Given the adverse perinatal outcomes associated with IPV, adapting criminal justice and healthcare responses to address the differing needs of perinatal victims is clearly warranted. Specifically, as perinatal women continue to reach out to the criminal justice system for help but have lower case-conviction rates, it is even more crucial to arm them with future protections at the outset: safety planning, education about their legal rights, and education about shelter services. Because perinatal women are less likely to disclose IPV to ED healthcare providers but have higher overall ED utilization, active IPV screening of pregnant women could yield important benefits for both the woman and her future child. Increased identification within the healthcare setting could lead to more safety planning as well as increased victim access to community resources, which can then be leveraged against further abuse.
In the state of Michigan, some felonies, including aggravated assault, felonious assault, and criminal sexual conduct, are not specifically tagged as IPV within the prosecutor's records and, thus, were not able to be identified for inclusion in the study population. In the year 2000, of the estimated 170 IPV-related felonies, our database included 110 (65%) of them. As mentioned, our operationalization of help seeking was limited to institutional help seeking, specifically police, prosecutor, and ED venues. We were missing information on the use of other community-based IPV resources, such as shelter services, mental health, civil court, and primary care settings. In addition, we did not have information about informal sources of help seeking, such as churches, friends, and family. However, the two systems we included are the primary ones that IPV victims turn to for help, according to a Bureau of Justice Statistics Special Report.39
This study offers a unique window into the connections among perinatal status, IPV, and help seeking within both the healthcare and criminal justice settings. Being pregnant or postpartum does not appear to outweigh the many other factors, previously identified in the literature, influencing rates of IPV help seeking (social support, abuse severity, material resources, relationship status), but it may lower the threshold of help seeking, with pregnant and postpartum women seeking help at lower rates of injury. Although no more likely to visit an ED for the known assault, perinatal women have greater ED utilization in the 6 months leading up a police-reported IPV assault than do nonperinatal abused women. Within the criminal justice system, perinatal women have overall lower conviction rates, however, perhaps indicating perinatal victims' declining motivation to prosecute the father of their child. Finally, age and race matter: Older and Caucasian perinatal women are less likely to call police and more likely to visit an ED than older Caucasian nonperinatal women.
This research was supported by NIJ grant 2006-WG-BX-0007 (K.V.R. and C.C.) and NIMH grant K23 MH64572 (K.V.R.) and by Borgess Research Institute. We thank Patricia Smith, Director, Violence Against Women Prevention Program, Michigan Department of Community Health; the Kalamazoo Assault Intervention Program; the Borgess and Bronson Emergency Department staff; and the many research assistants who worked on this project. This project builds on data collected during the MEDCIIN Projects, collected under a CDC cooperative agreement U17/CCU551067.
The authors have no conflicts of interest to report.