Introduction

In 1985, the Surgeon General identified intimate partner violence (IPV) as a public health problem of epidemic proportions responsible for 30% of female homicides (Campbell, 1998; McAfee, 2001). A problem crossing racial/ethnic boundaries and socioeconomic strata, it is one of the objectives of the Public Health Service's Healthy People 2000 and 2010 to reduce the rate of physical assault by current or former intimate partners (United States Department of Health & Human Serves (USDHHS), 1995, 2000). In 1994, the Violence Against Women Act was legislated providing national coordination of efforts to reduce IPV (Haywood & Haile-Mariam, 1999). This paper reviews the epidemiology, significance, identification and screening, and interventions for IPV. Implications for nursing are discussed, concluding with directions for further investigation.

Epidemiology

The objective in Healthy People 2010 is to "reduce the rate of physical assault by current or former intimate partners" to a rate of 4 per 1000 women 12 years and older (USDHHS, 2000). Although violence by intimates has decreased in the past decade, numbers remaining are substantial. In 1998, approximately 900,000 women reported robbery, physical, or sexual assault by intimates, reduced from 1.1 million in 1993 (Rennison & Welchans, 2000). IPV accounted for 22% of violent crimes against women between 1993 and 1998 (vs 3% for men) (Rennison & Welchans). The number of yearly IPV cases reported in past years ranges from 4% to 14% in population based studies and up to 44% in health care settings (Campbell, 2001; Jones et al., 1999; Wilt & Olson, 1996). Lifetime estimates vary from 5% to 51% (Bauer, Rodriguez, & Perez-Stable, 2000; Campbell, 2001; Schafer, Caetano, & Clark, 1998). Women are significantly more likely to be physically or sexually assaulted by a current or former intimate partner than an acquaintance, family member, friend, or stranger (Tjaden & Thoennes, 2000). As high as these figures are, it is commonly accepted that they (particularly crime data) represent underestimates of the true incidence and prevalence of IPV.

Lethality of IPV

Homicides by intimates have also decreased; however, the 30% rate of female victims killed by intimates has remained unchanged since collection of such statistics in 1976 (Rennison & Welchans, 2000). While there was a 4% decline of male victims between 1976 and 1998, there was only a 1% decline for female victims (Rennison & Welchans). Of all murders attributed to intimate partners, almost 75% involve female victims (Rennison & Welchans).

In 1997, homicide was the second leading cause of death among U.S. women aged 15 to 24 and the leading cause of death for African American women in the same age group (Centers for Disease Control (CDC), 2000). Most often the perpetrator identified is her current or ex-intimate partner (Campbell, 1994). In one state, homicide by an intimate partner accounted for more than a third (36.4%) of homicide-related maternal deaths (death within 1 year of birth or other pregnancy outcome) (Parsons & Harper, 1999). In this same study obstetricians were contacted to inquire about the woman's known history of IPV, substance abuse, or mental health problems. There was a known or suspected history of IPV in 2 of 5 cases of suicide; depression was known in 1 of 5 women who committed suicide. Among the violent deaths not attributed to homicide, 26.8% had a known history of substance abuse. Substance abuse is positively correlated with IPV including substance abuse by the victim (El-Bassel, Witte, Wada, Gilbert, & Wallace, 2001; Martin, Clark, Lynch & Kupper, 1999; Wingood, Diclemente, & Raj, 2000), perpetrator (Kub, Campbell, Rose, & Soeken, 1999; Sharps, Campbell, Campbell, Gary, & Webster, 2001), and both victim and abuser (Gilbert, El-Bassel, Rajah, Foleno & Frye, 2001; Muhajarine, 1999). The most common mental health consequence of IPV is depression (Campbell, 1994). Therefore, IPV was likely directly or indirectly related to more than the reported 36% of the homicide-related maternal deaths. In other research, prior domestic violence was present in approximately two thirds of homicide of women by intimate partners (Campbell, 1994; Morocco, Runyan, & Butts, 1998). Unfortunately, most health care providers do not know of their patients' domestic violence victimization, because they fail to make routine inquiry about this potentially lethal maternal health problem (Chamberlain & Perham-Hester, 2000).

Multiple investigators have determined risk factors which when present increase the risk of homicide (Campbell, 1994). These risk factors are summarized in Table 1. IPV is also associated with women killing their partners. Just as when the women are killed, most of these homicides are associated with prior IPV with the woman as the victim. Browne observed the following factors to be present in her sample of battered women who killed their partners: his substance use and frequency of intoxication, frequency of violent incidents, forced sexual acts, his threats to kill, her severity of injuries, and her threats of suicide (Campbell, 1994).

Substance abuse, particularly alcohol, while not commonly believed to be the cause, is significantly correlated with IPV. Abuse by male perpetrators appears to be a stronger correlate compared to female victims. Sharps and colleagues (2001) observed in their investigation of intimate partner femicide that perpetrator problem drinking was associated with an 8-fold increase with IPV and a 2-fold increase in femicide or attempted femicide. In contrast, most women (victims or their proxies) denied substance use during the most severe or lethal incident and it was not associated with femicide.

The Danger Assessment

(Campbell, 1994)developed a clinical instrument called The Danger Assessment which women can self-administer, designed to help them evaluate their own risk of homicide (See Table 2). The Danger Assessment also includes a question to evaluate women's suicide risk and one to assess violence toward her children. Depression may also be suggested and if present, indicates further evaluation. The Danger Assessment has no minimum score where lethality significantly increases; instead women must estimate risk based on their unique circumstances. By counting the days, the previous year's calendar assists women in determining whether the abuse has increased in severity or frequency. Women approximate the incident's duration and rank the incident on the scale included with the Danger Assessment. Investigations of the Danger Assessment document that when marking the calendar, up to 38% of women changed their answers from no to yes supporting patterns of increased frequency and severity of IPV (Campbell, 1994). This instrument may increase women's awareness of their lethal risk of abuse.

Risk Factors and Health Effects

IPV is comprised of three types of abuse, physical, sexual, or psychological abuse. They tend to overlap, with each violent relation rarely involving solely one type of abuse. Multiple investigators have determined risk factors that accompany IPV. Risk factors associated with IPV include age, marital status, and employment (Humphreys, Parker, & Campbell, 2001). Women between the ages of 16 and 24 years had the highest rates of IPV, currently 7.5 per 1,000 women (Rennison & Welchans, 2000). Women who are separated or divorced, whose male partner are unemployed, and have low combined income are also at greater risk for IPV (Humphreys et al.). This combination of socioeconomic risk factors implies that these women have less access to monetary and other resources and suggests that women without such risk factors and more resources have a greater capacity to prevent IPV or resolve it at very early stages. Additionally, women with these risk factors may be overrepresented in common data collection sites including the emergency department (ED), public clinics, and law enforcement agencies compared to women without these risk factors.

As a major public health problem, IPV has multiple negative health consequences for women. In addition to the physical injuries, such consequences include increased gynecological problems such as chronic pelvic pain and pelvic inflammatory disease, mental health problems such as depression, anxiety, sleep problems, and posttraumatic stress disorder (PTSD), musculoskeletal disorders such as back pain and headache, chronic gastrointestinal illnesses, and substance abuse among others (Campbell & Soeken, 1999; Coker, Smith, Bethea, King, & McKeown, 2000; Eby, Campbell, Sullivan, & Davidson, 1995; Leserman, Drossman, & Hu, 1998). Furthermore, violence continues during pregnancy for between 4% and 17% of women and is associated with increased rates of spontaneous abortion, preterm birth, and low birthweight (Gazmararian et al., 1996; Heath, 2001; Humphreys et al., 2001). It is therefore not surprising that women affected by IPV annually have increased use of health care services including emergency department and in-patient hospitalizations, and more contact with public health nurses and mental health professionals (Humphreys et al.; Liebschutz, Mulvey, & Samet, 1997; Sansone, Wiederman, & Sanson, 1997). Health care settings and nurses then are primary avenues for identification and intervention for female survivors of IPV.

Identification and Screening

Universal screening is advocated as an approach to increase identification partly because no demographic profile or pattern of injuries or clinical illness reliably identifies women affected by IPV. A major emphasis on reducing IPV rates is through timely identification of affected women seeking emergency care. While it may be assumed that battered women in the ED most often present with an injury from IPV, this may not be the case. In a study of 11 emergency departments in Pennsylvania and California, more than 33% of recently abused and 76% of women reporting current IPV stated they did not come to the ED for treatment of that injury (Glass, Dearwater, & Campbell, 2001). In another study, women visiting the ED were asked about their most common sources of help utilized (Pakiesar, Lenaghan, & Muelleman, 1998). The most common source of help actually used was family and friends followed by police and then the ED.

Unfortunately while most abused and non-abused women support universal screening, few women are screened. In the analysis by Glass and colleagues (2001), less than 25% of women were asked about IPV by the ED staff. Women presenting with acute trauma from abuse had higher (39%) screening rates compared to women reporting prior-year abuse (13%). Women reporting current abuse however, were significantly less likely to support routine screening (80% vs 89%), although the vast majority did. In contrast, in Gielen and colleague's study (2000) abused women were 1.5 times more likely than women without abuse to agree with universal screening. Similar to Glass et al.'s investigation, more abused (1.4 times more likely) women believed that reporting abuse to law enforcement should be the woman's decision.

Recognizing low rates of IPV identification, investigators have researched barriers to screening. Larkin and colleagues (Larkin, Hyman, Mathias, D'Amico, & MacLeod, 1999) identified patient and provider factors for female ED patients through a random medical chart review of 1,638 records where almost a third (29.5%) were screened for IPV. Screening varied with severity of the woman's condition, type of presenting complaint, and time of day. Women with nonpsychiatric, less acute complaints and women coming for care during daylight hours were more likely to be screened than women who presented with psychiatric, more acute complaints, or during the night shift.

Rodriguez (2001) examined factors associated with disclosure of IPV to physicians through telephone interviews with 375 culturally diverse women attending public clinics where 42% of the women stated they had communicated with a physician regarding their abuse. Direct physician questioning about abuse and African American ancestry were significant independent predictors of communication whereas immigrant status and patient concerns about confidentiality were associated with lack of communication. Women's perceptions that physicians did not ask directly about abuse, that they had insufficient time and interest in discussing abuse, fears about legal involvement, and concerns about confidentiality were barriers associated with lack of communication. Of all these factors, clinician inquiry was the strongest determinant of IPV disclosure.

Mandatory Reporting

There are a few states with laws mandating report of women with injuries resulting from IPV to local law enforcement agencies by health care workers; but it is only clearly mandated in all IPV cases, with or without serious injury, by the state of California (Campbell, 2001). Currently, no data exist to support the value of mandatory reporting in regard to it being beneficial or harmful to the women it was designed to protect; and California's level of reported abuse from health care sites has not changed since this legislation (Campbell, 2001). Although most women surveyed in ED settings supported mandatory reporting, abused women were significantly less likely to do so (76% vs 92%) (Campbell, 2001). Significantly more abused women believed that women's risk of abuse increased with mandatory reporting; and both abused and non-abused women believed that mandatory reporting would make women less likely to inform health care providers of the abuse. Rodriguez and associates (Rodriguez, McLoughlin, Bauer, Paredes, & Grumbach, 1999) examined California's physicians' perspectives on mandatory reporting and noted that approximately 59% of primary care and emergency physicians stated they may not comply with the reporting law if the woman objects. Primary care physicians reported lower compliance than emergency physicians and most physicians agreed that while mandatory reporting had benefits, it was not without potential risks to the woman's safety and subsequent concerns about violating medical ethics (e.g., overriding a mentally competent patient's decision).

Interventions

Interventions for IPV have been of two types: first, interventions to increase identification and screening of women and second, treatment of IPV. Many health care settings have implemented policies for routine screening and some have instituted training on how to screen for IPV. Overall, training to improve screening has been successful (although not 100%) at least in the short-term.

Thompson and associates (2000) conducted a randomized trial to improve case finding of domestic violence (DV) that included child abuse, in 5 primary care clinics. The intervention included strategic placement of posters and brochures, skills training for providers, and routine questions about DV on health questionnaires. Nine months after the intervention, providers believed they more often inquired about DV, had less fear of offending women, and fewer safety concerns. Case finding increased 1.3 fold and documented asking about DV increased 3.9 fold. Except for increased perceived asking, these results were also observed 21 months post intervention. Results obtained were primarily attributed to the presence of posters, brochures in the restrooms, and routine use of health questionnaires with the DV questions at physical examination visits. Environmental changes were important since abusers often accompany women to their health care visits. Wall posters portrayed valuable information without singling out a specific woman in the abuser's presence; and the restroom was the one location where he was not allowed to go.

McFarlane, Parker and colleagues (Parker, McFarlane, Soeken, Silva, & Reel, 1999; McFarlane, Soeken, Reel, Parker, & Silva, 1997) have tested an intervention to prevent further violence among pregnant women affected by IPV. Women in the intervention group were counseled about the cycle of violence, the Danger Assessment, available options, and resource referrals, and also empowered to make a safety plan. Women in the control group received a wallet-sized card with resource referrals but did not receive the counseling. Those in the intervention group had significantly less abuse at 6 and 12 months post-treatment and reported significantly more safety behaviors than women in the control group. The intervention is described fully in a March of Dimes publication that is easily available (McFarlane & Parker, 1994).

In another investigation, McFarland and Wiist (1997) examined the effect of an advocacy model involving lay persons referred to as "mentor mothers" for the purpose of establishing and maintaining contact with pregnant abused women until birth. Mentor mothers made a total of 922 contacts with 100 women, of which 74% were by telephone including 870 referrals for medical services, social services, abuse counseling, food, and other needs. This was a preliminary report and did not include outcomes specific to IPV.

In another investigation, women randomly assigned to a community-based advocacy intervention reported less violence, higher quality of life, and being more able to access community resources (Sullivan & Bybee, 1999). Advocates for this intervention were female undergraduate students who assisted abused women in making safety plans.

Nursing's Role

One of the most important things nurses can do to address IPV is to ask all women about violence. This must be done privately, and can occur during any patient encounter including during prevention-related visits (e.g., hospital admissions, annual exams, prenatal visits, sports physicals and pre-employment exams), trauma-related visits, and follow-up visits for chronic problems.

If obtaining privacy is difficult, as often women's partners accompany them into the exam room, nurses can ask questions in the restroom. Questioning should not be a one-time occurrence but at least annually or any time IPV is suspected. When asking questions, nurses should focus on behaviors and avoid the terms violence or abuse because women may not have the same interpretation. Examples of key questions that should be included in routine assessments are listed in Table 3. When a woman discloses a history of IPV, the nurse should validate her experiences and fear. Her immediate safety must be assessed and the woman should be directly asked if she feels safe going home. The Danger Assessment should also be given or self-administered to further evaluate safety, particularly for homicide risk. Nurses should be aware of laws in their jurisdiction regarding mandatory reporting and if reporting is required, let the woman know what must be done and why. If women are determined to be in danger, they should be assisted in accessing community resources such as shelters. They should be informed that the likelihood of abuse occurring again, escalating and increasing in severity is great. Women should accordingly be asked if they could tell when her partner is about to harm her prior to it occurring.

Nurses should assist women in creating a safety plan to be implemented as necessary (See Table 4). Most importantly, follow-up is critical and may be in the form of routine appointments or telephone calls. Nurses should document everything in the medical record including completion of a lethality assessment (See Table 5) (Campbell, 1994).

Currently, there is more emphasis on screening, referral and treating the consequences of IPV rather than its causes and prevention. There needs to be more coordination between and integration among health, social welfare, and criminal justice settings. This should occur not only for IPV but all forms of family violence including child abuse, elder abuse and violence by children against parents because IPV does not often occur in isolation. For example, IPV overlaps with child abuse in an estimated 40% to 60% of families (Gelles, 2000). Unfortunately, advocates for each population group often have different agendas that are at times contradictory. For example, pediatricians investigating possible cases of child abuse often overlook evidence of violence against the children's mothers (Gelles). Furthermore, child protective investigators do not routinely receive training in identification and response to IPV (Gelles). Finally, child abuse cases are settled in juvenile and family courts, while IPV cases are settled in district or criminal courts. Decisions made in juvenile courts may conflict with those made in criminal courts.

Advocates are used in the United Kingdom to be a support person and guide through the criminal justice system (Heath, 2001). This advocate is available in the ED setting and facilitates what may be women's initial disclosure of violence. Advocates such as these can be made available to women in the criminal justice systems to assist in accessing the health care system for problems related to IPV. A case management approach would be useful to follow not just women but the entire family or household. Training on screening and treatment should be required annually in not only health care, criminal justice and social welfare settings but also at all workplaces. Since nursing is primarily a female profession, it can be assumed that substantial numbers of them are also affected by IPV. Women in health care, social welfare, and criminal justice who are involved in IPV probably have special needs to be considered, since they are normally in the position of helper rather than being helped.

To date, much of the focus has been on identification, screening, and treatment of IPV in the ED setting with less emphasis in primary care settings where affected women likely spend more time. Recommendations from the Family Violence Prevention Fund (www.FVPF.org) include universal screening in all settings for all females over age 14 regardless of whether IPV is suspected (Saltzman, Green, Marks, & Thacker, 2000). McFarland and associates (1997) in their study of resource use by abused women, recommended that providers in primary care settings must become more involved by coordinating services for women rather than simply referring them.

Directions for Further Investigation

Comparatively little is known about prevention of IPV including information about male perpetrators and dating violence during the adolescent years. The primary response to male perpetrators has been to arrest and incarcerate with or without counseling. Analyses to date have not demonstrated broad reductions in IPV (Gelles, 2000). For abusers who were married or employed, incidence of IPV is decreased after an arrest. Arrest however, has no effect on abusers who were unmarried or unemployed (Gelles). For many men, first time arrests do not earn them jail time; instead they are directed to counseling or '˜batterers' groups (Gelles). To date, questionable evidence exists regarding the effectiveness of batterers' groups in decreasing violence. In fact, well-designed studies have not found reductions in violence resulting from treatment (Gelles). Even less is known about preventing the batter's violent behavior or terminating it in the early stages. Attention can be directed, however, to the adolescent years.

Dating violence has been documented in the middle adolescent years ranging from 10% to 25% of high school students (Wekerle & Wolfe, 1999). Gender patterns are less clear although girls are still more likely than boys to be physically injured. As a developmental period where principles of healthy intimate relationships are learned, it is a prime time to implement preventive interventions. School nurses could be particularly effective in that regard.

Conclusion

Nursing can be critical in the campaign to identify, intervene and reduce intimate partner violence. Professional organizations such as the ANA, Emergency Nurses' Association, Association for Women's Health, Obstetrical, and Neonatal Nursing, the American Association of Colleges of Nursing, and the American College of Nurse Midwives all support routine screening for IPV by nurses. There are two organizations, the Nursing Network on Violence Against Women International and the Forensic Nurses' Association that are specifically concerned with violence against women (Campbell, 1998). Nursing research is recognized as important in uncovering some of the complexities related to health care responses to IPV and testing interventions that can be implemented in health care settings (Humphreys et al., 2001). Even so, nurses are not yet routinely screening for IPV in every health care setting, and our tremendous potential to be a critical component in the quest to end IPV is not yet realized (Chamberlain & Perham-Hester, 2000; Limandri & Tilden, 1996).

Authors

Benita J. Walton-Moss, DNS, RNCS, FNP
E-mail - bmoss@son.jhmi.edu

Dr. Walton-Moss is assistant professor and coordinator for the family nurse practitioner program. She has 19 years of clinical practice much of it in women's health care. Her clinical practice is currently at Johns Hopkins Breast Center and the Family and Children's Center, a family practice clinic.

Jacquelyn C. Campbell, PhD, RN, FAAN
E-mail - jcampbell@son.jhmi.edu

Dr. Campbell is Anna D. Wolfe Endowed Professor and Associate Dean for PhD and Research Programs. She developed the Danger Assessment and is a world wide expert on intimate partner violence including extensive research and publications in this area.