f.“

.uan-u r.

 

.IvIrQIk'

f.

"4
‘ ‘

 

  
 

 

 

 

 

 

 
 

 

 

 

 

A... v:

u a. ._

 

 

 

 

 

 

THU???

 

 

 

Q
goo \
LIBRARY
Mlchlgan State
Unive rsl'ty
This is to certify that the
dissertation entitled

DOMESTIC VIOLENCE SCREENING DURING PREGNANCY:
EFFECTS OF PROVIDER EDUCATION AND PATIENT ETHNICITY

presented by

Deborah Lynn Shapiro

has been accepted towards fulfillment
of the requirements for

Ph 0 D 0 degree in Psychology

 

  

:.o professor

  

Date 9/11/01

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cz/CIFICJDateDuopss-pts

 

 

DOMESTIC VIOLENCE SCREENING DURING PREGNANCY:
EFFECTS OF PROVIDER EDUCATION AND PATIENT ETHNICITY

By

Deborah Lynn Shapiro

AN ABSTRACT OF A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Psychology
2001

Alytia A. Levendosky, Ph.D.

ABSTRACT

DOMESTIC VIOLENCE SCREENING DURING PREGNANCY: EFFECTS OF
PROVIDER EDUCATION AND PATIENT ETHNICITY

By

Deborah Lynn Shapiro

It is estimated that 25% of women in the United States have been physically
and/or sexually assaulted by an intimate partner in their lifetime (Tj aden & Thoennes,
1998). Pregnancy is often considered to be a risk factor for domestic violence (DV), with
prevalence estimates of DV during pregnancy ranging from 0.9 % to 20.1% (Gazmararian
et al., 1996). The past 15 years have seen increased attention paid to DV in the health care
system. Universal DV screening during pregnancy is recommended, yet research has
found that few Ob/Gyn providers routinely screen for abuse, except when it is suspected
(Horan, Chapin, Klein, Schmidt, & Schulkin, 1998). There are numerous potential
barriers to DV screening and disclosure, and a theoretical model was developed for the
purpose of this project to help explain the complex process. Of particular focus was the
potential impact of race/ethnicity on the screening process. A universal DV screening
program was instituted in a residency program Ob/Gyn clinic. Clinic staff attended a
training workshop on DV and screening methods. The program was evaluated through
medical chart reviews and provider surveys prior to and six months following this
workshop. Furthermore, clinic patients completed surveys six months into the screening
program. Physicians screened 46.9% of patients for DV over a six-month period. Results

indicate that having physicians screen for abuse in addition to nurses and social workers

does not increase disclosure. Over time, providers reported increased comfort and skill in
conducting the screening. Patient race was not found to affect disclosure, but did impact
the rate of screening, with Caucasian women screened more frequently than women of
color. The majority of patients did support the concept of universal DV screening and
most reported that they would disclose abuse if asked. The perceived barriers to effective
screening and disclosure differ between providers and patients, identifying a need for
changes in clinic procedures and more attention paid to the concerns of the patient.
Women disclosing abuse were referred for, appropriate follow-up intervention with a
clinic social worker. These results point to the need for similar programs to be
implemented in varied clinic settings, including private practice clinics. Long-term
follow-up is also necessary to more fully understand the complex screening and
disclosure process. Finally, universal screening and appropriate intervention can be an
effective component of more expansive societal prevention and intervention efforts in the

area of violence against women.

ACKNOWLEDGMENTS

This project was a significant part of my life for over two years, and it would not
have been completed without the cooperation, assistance, and support of many people.
First, I am grateful to the staff of the OB/GYN clinic for allowing me to enter their
facility and assisting in the implementation of the project. When people are feeling
overworked already, it takes a special group of people to agree to what seems like
additional work. The patients in the clinic also deserve recognition, for contributing their
voices to the development of future DV screening programs.

The incredible and never-ending professional and personal support that I have
received from Dr. Alytia Levendosky throughout my graduate school career is beyond the
thanks that I can put into words. She has been a mentor, research advisor, teacher, and
friend — always seeming to know which role to take on at any given time. Thank you,
Alytia. The members of my dissertation committee - Drs. Bogat, Vasilenko, Sullivan,
and von Eye — each assisted with this project in different ways and helped see it to
fruition. I would also like to thank my team of research assistants - Sarah, Reena, Katie,
and Don — for all of the work they did throughout the year that I could not possibly have
done on my own.

Finally, those people (you know who you are) who did not work alongside me on
this project, but provided me with much-needed support and encouragement at the end of
the day, were also vital in seeing that the project was completed. Thank you for never
doubting that I could actually do it, for listening when I was frustrated, and for

celebrating with me when I achieved success.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
INTRODUCTION ............................................................................................................... 1
Risk of Domestic Violence during Pregnancy ......................................................... 3
Domestic Violence during Pregnancy: Prevalence .................................................. 4
Domestic Violence during Pregnancy: Correlates and Risk Factors ....................... 5
Pregnancy Outcome ................................................................................................. 7

The Role of the Health Care Provider .................................................................... 11
Screening for DV: Intervention and Prevention ................................... - ................. 14
Barriers to DV Screening ....................................................................................... 18

The Barriers Model ............ -. ....................................................................... 18

Hierarchical Model of Leisure Constraints ................................................ 20

Physician Barriers to DV Screening .......................................................... 21

Patient Barriers to disclosure of DV .......................................................... 23

Integrated Model of Barriers to Effective DV Screening .......................... 25

Cultural Issues and Domestic Violence ................................................................. 30
African American Women ......................................................................... 33

Hispanic Women ........................................................................................ 36

Asian and Asian-American Women .......................................................... 39

Immigrant Women ..................................................................................... 42

Conclusions ............................................................................................................ 44
HYPOTHESES .................................................................................................................. 47
METHOD .......................................................................................................................... 49
Setting .................................................................................................................... 49
Procedures .............................................................................................................. 50
Preliminary Activities ................................................................................ 52

Archival Data Collection ........................................................................... 53

Provider Training ............... ' ........................................................................ 56

Patient Surveys ........................................................................................... 59

Intervention ................................................................................................ 61

Measures ................................................................................................................ 61
Patient Survey ............................................................................................ 61

Provider Survey ......................................................................................... 62

Abuse Assessment Screen .......................................................................... 64

RESULTS .......................................................................................................................... 66
Archival Data Collection ....................................................................................... 66

Initial Chart Review ................................................................................... 66

Second Chart Review ................................................................................. 67

Provider Training ................................................................................................... 72
Survey of Patients .................................................................................................. 77
Barriers to Disclosing Domestic Violence ................................................. 83

Intervention ........................................................................................... . ................. 86
DISCUSSION .................................................................................................................... 88
Archival Data Collection ............... ° ........................................................................ 88
Provider Training ................................................................................................... 95

Patient Survey ........................................................................................................ 97
Intervention .......................................................................................................... 103
Limitations ........................................................................................................... 104
Implementation Issues ......................................................................................... 106

Future Directions ................................................................................................. 107
APPENDIX A: Provider Training Outline ...................................................................... 110
APPENDIX B: Patient Survey ......................................................................................... 118
APPENDIX C: Provider Pre—Training Survey ................................................................ 122
APPENDIX D: Provider Post-Training Survey ............................................................... 128
REFERENCES ................................................................................................. . .............. 136

vi

LIST OF TABLES

Table 1. Project Tirneline ................................................................................................... 51
Table 2. Cronbach’s Alpha Reliability Coefficients for Provider Survey Factors ............ 64
Table 3. Number of DV Screenings by Patient Race/Ethnicity ......................................... 71
Table 4. Patients’ Definitions of Domestic Violence ........................................................ 78
Table 5. Barriers to Disclosing Domestic Violence to a Physician ................................... 85

vii

LIST OF FIGURES

Figure 1. Potential barriers to domestic violence screening and disclosure in the health
care setting ......................................................................................................................... 26

Figure 2. Direct and indirect effects of training on provider barriers to domestic violence
screening ............................................................................................................................ 74

Figure 3. Patients’ comfort in discussing DV with their physicians .................................. 79

. viii

INTRODUCTION

According to the National Violence Against Women Survey (see Tjaden &
Thoennes, 1998), 25% of women in the United States have been raped and/or physically
assaulted by a current or former spouse, cohabitating or dating partner in their lifetime.
Almost 8% of women reported sexual assault by an intimate partner, while 22% reported
physical assault. It is important to note that while the rates of victimization increase when
assaults from acquaintances or strangers are included, the vast majority (76%) of violence
perpetrated against women is by an intimate partner. Domestic violence (DV) is not a
problem limited to this country; internationally, one-fifth to over one-half of women
interviewed in countries across the world reported physical abuse by a male partner
(Heise, 1994). Because of varying definitions of domestic violence across cultures, it is
difficult to estimate worldwide rates that include psychological, verbal, and other non-
physical forms of abuse.

In 1996, a statewide survey on violence in the lives of women in Michigan was
conducted (Michigan Department of Community Health, 1996). According to this survey,
58% of Michigan women have experienced some type of violence by a man since age 16,
with 38% reporting physical violence and 40% reporting sexual violence. Thirty-eight
percent (3 8%) of women who had ever had a partner had experienced some type of
partner violence. These rates emphasize the need for heightened prevention and
intervention efforts in the area of violence against women (V AW).

Recently, medical professionals have begun to address the issue of domestic

violence in their patients’ lives, including it as a women’s health issue. Health care

providers can play a vital role in the prevention of domestic violence and intervention
with women who are in abusive relationships. One particular area of recent interest has
been the occurrence of DV during pregnancy. There may be increased risk for violence
during pregnancy (Stewart & Cecutti, 1993), and there are specific health effects that
might occur as a result of DV during pregnancy. Furthermore, the prenatal period is a
time when a woman ofien meets with her physician on a regular basis and a relationship
of trust can develop, presenting an ideal time for prevention and intervention.
Obstetricians need to be aware of both the potential effects that abuse might have on the
pregnancy, as well as how to help a patient who may be experiencing domestic violence.

There have been a number of advances in the health care field concerning
violence in patients’ lives in the past 10 years, and many providers now receive DV
education and are discussing DV with their patients. However, the education that health
care providers receive is often concerning DV in a “typical” patient, ignoring factors that
might be specific to women of different socioeconomic, racial, or ethnic backgrounds.
Although women’s experiences of violence are remarkably similar across cultures, how
to intervene as a health care provider may vary in a number of ways.

The following review will provide a background of the risk, prevalence,
correlates, and effects of DV during pregnancy, followed by a summary of the role that
health care providers can assume in the prevention and intervention of domestic violence,
starting with universal DV screening in their patient population. The potential barriers to
effective DV screening and intervention are summarized, emphasizing the need to
remove barriers prior to implementing new screening protocols, through training of health

care staff. A model representing these barriers is also presented. A discussion of cultural

issues related to DV follows, indicating the need for more thorough training for providers
who work with an ethnically diverse population.

Risk of Domestic Violence during Pregnancy

 

Pregnancy is often considered to be a risk factor for domestic violence. It may be
the time when physical abuse first occurs within a relationship, or ongoing abuse may
escalate during the pregnancy. Women abused prior to pregnancy are more likely to be
abused during pregnancy, with about half of all battered women sufi‘ering abuse during
pregnancy (Bash & Jones, 1994; Bohn, 1990).

A number of reasons have been cited to explain battering during pregnancy. These
reasons fall into four general categories: 1) abuse directed at the women related to her
increased vulnerability and defenselessness (Berenson, Stiglich, Wilkinson, & Anderson,
1991; Stewart & Cecutti, 1993; Campbell,'Oliver, & Bullock, 1993); 2) anger and/or
jealousy towards the fetus, who may be seen as competition or as interfering with the
abuser’s control (Bash & Jones, 1994; Berenson et a1, 1991; Campbell, Oliver, &
Bullock, 1993); 3) stress of the family transition, including economic pressures and
sexual frustration (Bash & Jones, 1994; Berenson et al., 1991; Stewart & Cecutti, 1993);
and 4) continuation of previous abuse, or “business as usual” (Campbell, Oliver, &
Bullock, 1993).

Women in physically abusive relationships are also at a higher risk for coercive
and forced sex. It is estimated that one-fourth to more than half of all battered women are
also sexually assaulted by their partners, which has important health implications during
pregnancy and in the postpartum period (Bohn, 1990; Campbell, Oliver, & Bullock,

1993). At times, pregnancy is actually a result of abuse. Domestic violence can lead to

unwanted pregnancy through direct or indirect means; directly through rape or the
inability to effectively use contraception, or indirectly through risk behaviors such as
alcohol use, early sexual initiation, or unprotected sex (Heise, 1994). Not only do the
physical effects of such abuse need to be considered, but the emotional impact on a
woman who has been assaulted must also be taken into account, during and afier
pregnancy (Mezey & Bewley, 1997). The increased risk for DV during pregnancy is also
illustrated in the prevalence rates, which exceed rates of other relatively common
pregnancy complications.

Domestic Violence during Pregnancy: Prevalence

 

In a review of the literature, Gazmararian et a1. (1996) found the prevalence of
violence during pregnancy to be between 0.9% and 20.1%. These rates were affected by
the methods in which women were questioned about their experiences of violence,
including the number of times they were asked and who asked. While the majority of
studies included in the review reported rates between 3.9% and 8.3%, the prevalence rates
were greatest (16% - 20.1%) in studies when assessment occurred more than once, in
person, and continued later in the pregnancy. Thus, it appears that a more thorough
assessment of intimate violence during pregnancy, conducted by a trained interviewer,
will lead to higher rates of reporting.

Reporting rates also vary depending on one’s definition of domestic violence.
Among one sample of low-income pregnant women, 65% experienced either verbal or
physical abuse during their pregnancies (O’Campo, Gielen, Faden, & Kass, 1994). This

rate is much higher than when considering physical abuse alone. It is likely that the

majority of women reporting physical abuse during their pregnancy are victims of
psychological or verbal abuse as well (Stewart & Cecutti, 1993).

Some studies have found that the reporting rates of domestic violence during
pregnancy increase in the third trimester. Webster, Sweett, and Stolz (1994) reported an
increase in violence during the current pregnancy from 5.8% of patients at the initial visit
to 8.9% at 36 weeks. Almost 24% of those women reporting abuse stated that the abuse
had increased since confirmation of the pregnancy, meaning this increase was not solely
due to increased reporting. In a second study, 19% of pregnant women reported
experiencing moderate or severe abuse prenatally, while 25% reported such abuse
postpartum, emphasizing the need for continued screening and care throughout and
following the pregnancy (Gielen, O’Campo, F aden, Kass, & Xue, 1994). These findings
also illustrate the need to exercise caution when interpreting prevalence estimates in
research studies. Abuse reporting can be affected by interviewer characteristics, specific
questions that are asked, when screening occurs, and definitions of DV. Additionally,
estimated prevalence rates can be confounded in research studies by a variety of common
correlates.

Domestic Violence during Pregnancy: Correlates and Risk Factors

 

A large number of correlates of and risk factors for domestic violence during
pregnancy have been identified. Some of the correlates found across studies include being
younger (Cokkinides & Coker, 1998; Gazmararian et al., 1995; O’Campo et al., 1994),
being unmarried, receiving social assistance and/or living below the poverty line, having
an unintended or unwanted pregnancy (Cokkinides & Coker, 1998; Gazmararian et al.,

1995; Stewart & Cecutti, 1993), having less than a high school education (Cokkinides &

Coker, 1998; Stewart & Cecutti, 1993), drug, alcohol and/or tobacco use (Campbell,
Poland, Waller, & Ager, 1992; Cokkinides & Coker, 1998; Stewart & Cecutti, 1993),
having higher parity [more pregnancies] (Cokkinides & Coker, 1998; Gazmararian et al.,
1995), receiving inadequate prenatal care, stressful life events including housing
problems, unemployment, or a lack of social support (Campbell, Poland, Waller, & Ager,
1992; Cokkinides & Coker, 1998), and psychological distress including anxiety and
depression (Campbell et al., 1992; Stewart & Cecutti, 1993).

Studies examining differences in experiences of abuse across race have produced
inconsistent findings. Race is often not associated with physical violence when
controlling for age, poverty, and other stressful life events (Cokkinides & Coker, 1998).
Berenson et a1. (1991) explained that women of different races present different profiles
in terms of risk factors, such as marital status, parity, and substance use. In other words,
some characteristics may be identifying risk factors for violence for women of one racial
group, but not for those of another. For example, being divorced or separated is a
correlate of DV for white non-Hispanic and black women, but not for Hispanic women.
Overall, there is no standard description of a battered woman that can be used to identify
a woman as such, providing more evidence for the need for routine screening and direct
questioning. .

Adolescence appears to be a particularly high risk factor for abuse during
pregnancy, with as many as 43.7% of pregnant adolescents reporting physical abuse
(Webster, Sweett, & Stolz, 1994). Adolescents are at risk for abuse not only from their
romantic partner, but from members of their family of origin as well. Berenson, San

Miguel, and Wilkinson (1992) found 40% of adolescents reporting physical or sexual

abuse were victimized during pregnancy. Although there appeared to be more abuse from
members of family of origin than partner overall, partners were responsible for more of
the battering during pregnancy. It is interesting to note that while the risk of violence
might be greater for pregnant adolescents than adults, abused adult women have been
found to score higher on danger assessment screens and scales of severity of abuse
(McFarlane & Parker, 1994). This finding may actually make identification of an abused
adolescent more difficult, due to fewer physical injuries.

Adolescents who become pregnant are also at greater risk for having experienced
sexual abuse as children (Berenson, San Miguel, & Wilkinson, 1992). Some adolescents
may become pregnant as a result of sexual assault, with 23% of teens in one study who
experienced unwanted sex becoming pregnant by the perpetrator (Heise, 1994).
Adolescent patients may not be willing to disclose such information voluntarily, so it is
the health care provider’s responsibility tobe aware of correlates of teen pregnancy, such
as abuse. The importance of this awareness becomes clearer when considering the
potential adverse pregnancy outcomes, discussed below.

Pregnancy Outcome

 

There have been variable findings in the research literature concerning the effects
of physical abuse on pregnancy outcome. One reason that DV brings potential risk to the
health of the woman and the fetus is that the abdomen and reproductive organs are
frequently targets of abuse. In fact, some studies have found the abdomen to be the most
common area struck during pregnancy, being hit twice as frequently in pregnant women
compared to women battered when they were not pregnant (Berenson et al., 1991;

Stewart & Cecutti, 1993). Physical trauma, intentional or otherwise, to the maternal

abdomen can cause adverse pregnancy outcomes (Petersen et al., 1997). Abdominal
trauma may cause fetal injury or death through direct trauma to the fetus or indirectly
through uterine trauma, sometimes resulting in miscarriage (Bohn & Holz, 1996).

It has also been proposed that the repetitive nature of domestic violence may
explain the increased risk (compared to other forms of abdominal trauma) for fetal
demise, along with the occurrence of assaults during the third trimester and direct blows
to the abdomen (Ribe, Teggatz, & Harvey, 1993). Thus, it appears that accidental injury
to the abdomen may not result in the same consequences as abdominal injury resulting
from intentional abuse.

An additional factor that can lead to prenatal health problems is late entry to care
or inconsistent medical check-ups, which are behaviors associated with domestic violence
(Mayer & Liebschutz, 1998). Late entry to prenatal care is a risk factor for low birth
weight, as well as other risks to infant and maternal health (Mayer & Liebschutz, 1998).
Batterers may prevent women from entering prenatal care in the first trimester or meeting
consistently with their providers because they feel threatened by the woman developing a
relationship with her obstetrician. Women may also miss appointments because they do
not want their doctors to see physical signs of abuse.

A review of the literature examining outcomes related to violence during
pregnancy found no consistent outcome associated with violence, possibly due to
methodological limitations and differences between reviewed studies (Petersen et al.,
1997). Only mean birthweight and incidence of low birthweight (LBW) were found to be
statistically different between abused and nonabused women in more than one study. In

one study, battered women were two to four times more likely to give birth to low birth

weight infants than nonabused women (Bullock & McFarlane, 1989). LBW was
associated with a history of physical abuse regardless of race, smoking, alcohol use,
prenatal care, prior abortions, maternal complications, and hospital setting (public vs.
private).

Other negative effects of DV on pregnancy have been found, although not as
consistently. Berenson, Wiemann, Wilkinson, Jones, & Anderson (1994) found that
women who experienced physical abuse during pregnancy were twice as likely to have
preterm labor and chorioamnionitis than nonabused women were. In a sample of low-
income women who reported physical abuse during pregnancy, there was a greater
chance of fetal distress or fetal death, and their infants had to be kept hospitalized longer,
even after controlling for smoking and maternal age (Dye, Tolliver, Lee, & Kenney,
1995). Although evidence for infant mortality following abuse is inconsistent in the
literature, any incidents should be cause for concern. It is an area that requires further
study to discern the probability and mechanisms by which fetal demise might occur
following a domestic assault.

The inconsistent findings in the research literature concerning effects of DV on
pregnancy outcome may partly be due to the confounding influence of other factors, such
as socioeconomic status (SES) and substance use. The risk of adverse pregnancy outcome
increases with lower SES, alcohol use, and other substance use, which are also correlates
of violence during pregnancy. Thus, the adverse effects of domestic violence on
pregnancy outcome may be difficult to detect (O’Campo et al., 1994).

Physical abuse might directly lead to such outcomes as LBW or preterm delivery,

and/or indirectly affect behavior and/or physiological responses which cause adverse

9

pregnancy outcomes (Petersen et al., 1997). For example, stress related to ongoing
domestic violence may affect a woman’s self-care during pregnancy and is also
associated with such behaviors as smoking and substance abuse, which are associated
with adverse birth outcomes. Abused women are more likely to smoke, use alcohol and
other drugs during pregnancy, are less likely to receive adequate prenatal care, and are
more likely to have unhealthy diet and poor weight gain (Bohn & Holz, 1996). These
unhealthy behaviors may represent poor coping strategies, a perceived lack of control
over the health of the fetus, the greater rate of unwanted or unintended pregnancy, or be a
result of the batterer controlling access to health care (Bohn & Holz, 1996). '

In one of the few studies to examine psychosocial stressors along with physical
symptomatology, Dye et a1. (1995) found that women abused during pregnancy had more
frequent stress-related physical problems, and they reported more tension and arguments
with their partner since the pregnancy began. They were more likely to have separated
from their partner and to be homeless, and they were also more likely to smoke and drink
alcohol during pregnancy. They also reported more worrying, negative attitudes, and
dissatisfaction about different aspects of the pregnancy.

Many women who are physically abused by their partners are also sexually
abused by them, and they may have also suffered childhood sexual abuse (Bohn & Holz,
1996). The effects of these different forms of abuse are often the same, including stress-
related symptoms and more direct physical problems. Sexual assault can cause a variety
of pregnancy complications and health problems prenatally and in the postpartum period
(Bohn & Parker, 1993). For example, sexual assault can increase the risk of miscarriage

or stillbirth, preterm labor, complicated deliveries, or breast-feeding problems. Sexual

10

intercourse in the early postpartum period can result in endometriosis or inhibit healing of
episiotomies or lacerations. Often related to sexual assault, sexually transmitted diseases
(STDs) are also commonly reported by battered women, and can be linked to many health
problems in pregnancy, including premature membrane rupture, preterm birth, and
chorioamnionitis (Bohn & Parker, 1993). STDs can also result in complications such as
endometriosis, genital and peritoneal infections, infertility, and ectopic pregnancy.

Not only does abuse negatively impact both a pregnant woman and the fetus, but
DV during pregnancy has also been linked to maternal mortality. Intentional injury,
including homicide and suicide, have been found to be a significant cause of maternal
mortality cross-culturally (Heise, 1994). Studies across the country have found that more
pregnant women are killed by intimate partners than by any single medical complication
of pregnancy (Mayer & Liebschutz, 1998). Homicide, most frequently committed by an
intimate partner, is the leading cause of injury-related death among pregnant women in
US. urban areas (Dannenberg et al., 1995). Among these victims, women ages 15 - 34
and black women are at particularly high risk for homicide, compared to other pregnant
women and women in the general population.

When considering the numerous potential pregnancy complications related to DV,
the importance of the health care provider’s involvement becomes more apparent. Many
of these complications and outcomes might be prevented, or at least treated, if the
provider is aware of the presence of DV in the woman’s life.

The Role of the Health Care Provider
A significant portion of the work on VAW prevention in the health care setting

has been initiated and undertaken by nurses and nurse-midwives. However, an increase in

.11

the efforts of other health care providers has more recently begun. Since the Surgeon
General’s workshop on violence in 1985, medical associations such as the American
College of Obstetricians and Gynecologists (ACOG), the American Medical Association
(AMA), the American Medical Women’s Association (AMWA), the American Academy
of Family Practice (AAF P), and others have led educational campaigns on domestic
violence (Flitcrafi, 1993). However, the lack of institutional change and individual
support has kept the issue of intimate partner violence from being fully incorporated into
medical training and practice. Secondary prevention efforts, such as routine screening and
assessment, need to be incorporated into the changing medical environment, in a similar
way to other public health issues such as smoking cessation and HIV prevention
(Flitcrafi, 1993). Yet these efforts must be more comprehensive, incorporating primary,
secondary, and tertiary interventions. The health care provider’s role in coping with and
preventing domestic violence should include comprehensive assessment, collaboration
among health care professionals and other agencies, intervention, and prevention
(Kilpatrick, Resnick, & Aciemo, 1997).

The Joint Commission on the Accreditation of Healthcare Organizations
(J CAHO) 1992 guidelines for the treatment of adult and child victims of alleged or
suspected abuse or neglect require comprehensive identification, documentation,
treatment, and referral procedures, as well as ongoing training of health care providers, in
emergency departments and ambulatory care settings, including prenatal clinics (see
Sheridan & Taylor, 1993). The past six years have seen a number of changes in
emergency department procedures following these guidelines, but prenatal clinics have

not received the same amount of attention. One potential reason for the difference in

12

attention to DV between settings is the patient’s reason for seeking services. For example,
in an emergency department, a woman presents with injuries and her physician should
inquire if they are due to domestic violence. In contrast, if a woman is attending a regular
prenatal appointment and does not present with injuries, her physician may see no reason
to screen for domestic violence.

By integrating public health principles into medical practice, health care providers
can not only identify victims but help create effective primary prevention as well
(Rosenberg, F enley, Johnson, & Short, 1997). The prevention of VAW involves exposing
the abuse of individual women through assessment, as well as exposing the issue of
intimate violence to the public and to the health care community (King et al., 1993). Once
abuse is disclosed, important secondary intervention steps are listening, assessing safety,
validating her experience, providing advocacy and referral, mobilizing social support, and
documenting the abuse in her medical record (King et al., 1993.).

Obstetricians-gynecologists may be the primary care physicians for many women,
emphasizing the need for these specialists to be particularly aware of, and to assess for,
domestic violence. The importance of intimate violence within this health care
environment has become increasingly recognized and emphasized by the health care
system, indicated by such publications as the ACOG technical bulletin on domestic
violence (ACOG, 1995).

As was discussed above, abuse during pregnancy can lead to many complications
and negative health effects for both the woman and the fetus. Obstetricians routinely
screen for other health complications, such as high blood pressure and diabetes, although

their incidence is relatively rare. Abuse is a more common pregnancy complication than

13

hypertension, preeclampsia, hyperthyroidism, or gestational diabetes, yet it is rarely
screened for and often goes undetected (Bash & Jones, 1994). Physicians should be
trained to recognize common risk factors and “symptoms” of abuse and how to provide
referrals to social service agencies (Bash & Jones, 1994). However, they must also be
trained to recognize the limits of their role as physician, acknowledging that they cannot
force the woman to make changes in her life, or make decisions for her concerning
seeking services. Furthermore, they need to recognize that each woman’s experience of
being abused will be different, and to respect any resistance to discussing abuse that
women may display. Some patients may not disclose abuse the first few times they are
asked, but they will be more likely to discuss DV with their physicians if screening
occurs repeatedly (McFarlane, Parker, Soeken, & Bullock, 1992), and if there is a
relationship of openness and respect.

A number of hospital-based programs for domestic violence have been developed
in the past decade, and there are now a number of published protocols available to use as
models (see Domestic Violence Project, 1994; McFarlane & Parker, 1994; Sheridan &
Taylor, 1993; Warshaw & Ganley, 1995). While hospital-wide policy changes may take
some time, individual providers can begin prevention efforts by instituting procedures in
their own practice for assessment and intervention.

Screening for DV: Intervention and Prevention

 

Comprehensive, standardized assessment for domestic violence in prenatal
patients can be instituted as the standard of care without disrupting clinic flow. The use of
a structured assessment instrument has been found to improve detection of physical abuse

before and during pregnancy, increasing the opportunity to intervene (Norton, Peipert,

l4

Zierler, Lima, & Hume, 1995). However, such screening does not occur in most prenatal
settings. In a group of pregnant and non-pregnant patients in an urgent care ob/gyn clinic,
only 18% recalled ever being asked about abuse by a health care provider (McGrath,
Hogan, & Peipert, 1998). Of those women reporting a history of abuse, only 19% had
been screened by a health care provider. Of those reporting recent abuse, only 29% had
ever been screened. A second study reported that 92% of women did not discuss their
experiences of DV with their physicians (see Warshaw & Ganley, 1995). In one family
practice clinic, only six out of 394 women’surveyed had ever been asked by their
physician about DV, but 38.8% of the women had been physically abused by a partner
(Warshaw & Ganley, 1995).

In contrast to many health care providers’ beliefs that their patients would be
uncomfortable being asked about victimization, most female patients favor routine
assessment of physical and sexual assault, and report that they would answer honestly if
their physicians asked them about abuse history (Aciemo, Resnick, & Kilpatrick, 1997;
Friedman, Samet, Roberts, Hudlin, & Hans, 1992). Women who have been abused ofien
report being relieved and pleased that practitioners are being open about domestic
violence issues, and nonabused women often report knowing someone who is abused and
appreciate being given information about DV (McFarlane & Parker, 1994). Among a
group of women interviewed with a comprehensive psychosocial assessment, most not
only felt comfortable answering questions, but also felt validated and appreciated the
opportunity to reflect on their lives (Reid et al., 1998). Providers were initially reluctant
to uncover a history of abuse and risk losing control of the situation, but later reported no

unmanageable situations and increased comfort with the assessment. An added benefit

15

was the increased rapport reported by both providers and patients, creating an atmosphere
where information such as the experience of intimate violence could be disclosed.

The method of assessment may affect how comfortable women are disclosing
abuse in a health care setting. Thirty-six percent (36%) of a sample of DV victims in an
Emergency Department (ED) reported they would only disclose the violence if asked
directly (Hayden, Barton, & Hayden, 1997). About one-half of the women responding
stated they would feel comfortable disclosing abuse to an ED physician. An important
finding of this study was that while many women indicated they would not volunteer
information about DV, they reportedly would answer honestly if asked directly about DV.
Some women may not voluntarily disclose abuse, believing that their providers should
initiate discussion (Molliconi & Runyan, 1996).

A survey of ACOG Fellows indicated that most were aware of basic domestic
violence issues, most (71%) felt very or somewhat confident about their skills in
screening patients for DV, and 85% reported that their patients are not adverse to being
questioned about abuse (Horan, Chapin, Klein, Schmidt, & Schulkin, 1998). However,
68% of physicians responding only screened their pregnant patients when they suspected
abuse, versus routine screening. While 39% responded that they do routinely screen for
abuse at the first prenatal visit, only 3-5% screen their patients afler the first trimester. In
contrast, the majority of respondents routinely screen for tobacco or alcohol use, diet and
exercise, use of seat belts, and stress and depression.

It is surprising that despite patients’ support of regular abuse assessment, and the
preponderance of evidence of effects of violence on health, most physicians do not feel

that routine victimization screening should occur (Aciemo, Resnick, & Kilpatrick, 1997).

16

They may feel that their patients will discuss domestic violence with them if they so
choose. However, the evidence provided by female patients as discussed above indicates
that many opportunities to intervene are lost if DV assessment does not occur directly.

Health care providers do not need to assume all responsibility for intervening with
patients who are being abused, but they do need to play a role in identification and
referral. As Heise (1994) stated, “Perhaps the easiest and most significant step that
providers can take is to integrate screening for abuse into routine gynecological and
prenatal care” (p. 226). Health care providers can then refer women to other individuals
and agencies for further intervention. Providers can serve as important sources of
information and resources for women who are being battered, particularly when many
women are unaware of community resources (Helton & Snodgrass, 1987). Physicians and
nurses can serve an important role as the first contact for women who wish to seek help
regarding an abusive relationship.

A number of brief screening instruments have been developed and found to be
effective in identifying women experiencing DV (see Campbell, Harris, & Lee, 1995;
Norton, et al., 1995). Disclosure is generally maximized if the questions are asked at the
initial and all subsequent prenatal visits, in a private interview (Campbell, Harris, & Lee,
1995; King, 1993). Screening should also Continue into the postpartum period, since
research has found that the risk of moderate to severe violence is greatest during this time
(Mezey & Bewley, 1997). Even with the evidence supporting the importance of DV
screening and indicating patient support, there may be barriers that prevent DV screening

from being effectively carried out. Understanding these barriers may help to minimize

17

their impact. These barriers should be considered when designing DV screening protocols
to facilitate implementation of screening procedures.

Barriers to DV Screening

 

The question of why DV screening, may not be successfully implemented can be
answered by examining the potential barriers to effective screening and disclosure. By
summarizing the empirical research on barriers to DV screening and borrowing
theoretical models from other lines of research, a theoretical model of barriers to DV
screening and disclosure was developed. Within this model, one can more easily
understand how these barriers may operate as well as how to begin to minimize their
potential to occur. This proposed model was developed by incorporating The Barriers
Model for intervention with battered women (Grigsby & Hartman, 1997) with Crawford,
Jackson, and Godbey’s (1991) Hierarchical Model of Leisure Constraints. The model also
shares similarities with Gerbert et al.’s (1996) Systems Model. However, theBarriers
Model developed here allows for patient and provider barriers to be examined
simultaneously, using a slightly different conceptualization that was considered to be a
better theoretical fit for this study.

The Barriers Model. Grigsby and Hartman’s (1997) Barriers Model represents the
layers in a battered woman’s experience that may impede her safety and prevent her fiom
seeking help. This theoretical model was developed in response to intervention theories
and strategies that were victim blaming and pathologizing towards battered women and is
based on evidence from practice, but has not been empirically tested. The Barriers Model
consists of four layers: 1) barriers in the environment; 2) family and social role

expectations; 3) psychological consequences of abuse; and 4) childhood abuse and

18

neglect issues. Level 1 barriers include the batterer, money, transportation, police
response, criminal justice system, mental health system, physical and cultural
accessibility to shelters/services, language, culture, and immigration. Different aspects of
these environmental factors can impede a woman’s ability to seek assistance while in a
violent relationship. Level 2 barriers include effects of socialization such as values/beliefs
about relationships and abuse, identity, religious values, and family of origin
values/beliefs. Some women hold values that do not support her choice to end an abusive
relationship, which may be further impacted by a family of origin supporting more
traditional family roles. This level may be particularly impacted by traditional cultural
values. Level 3 includes defense mechanisms, physical/somatic results of abuse,
psychological consequences, isolation, brainwashing, and posttraumatic stress. The
defenses that have enabled the victim to survive in the abusive relationship may act
against her by impairing her ability to perceive danger, reducing her self-esteem and
mental capacity to choose to leave, and minimizing her resources. Finally, Level 4
barriers include the consequences of childhood abuse and neglect, impairing a woman’s
judgment in potentially dangerous situations and long-tenn psychological effects such as
depression, substance abuse, anxiety, and posttraumatic stress responses. While some of
the barriers in this model are within the victim herself, the primary focus is on external
barriers and resources. In this model, societal level barriers must be addressed before
individual change can be effective. By emphasizing the societal context of battering
rather than placing the primary focus on the individual, this model helps to explain the

barriers to leaving an abusive partner without placing sole responsibility on the victim.

19

Hierarchical Model of Leisure Constraints. The purpose of this model is to
understand barriers to participation in leisure activities. Although the topic of interest
may differ, it is a useful theoretical model that can help explain barriers to effective DV
screening. In the same way that individuals may face barriers to engaging in leisure
activities, they may face barriers in carrying out other behaviors, such as DV screening or
disclosure. In the Model of Leisure Constraints, participants are viewed as having to
negotiate a sequential, hierarchical series of constraints including intrapersonal,
interpersonal, and structural levels (Crawford, Jackson, & Godbey, 1991). Intrapersonal
barriers are those individual characteristics affecting leisure preferences. These can
include stress, depression, anxiety, attitudes, prior socialization, perceived self-skill, and
subjective assessment of the appropriateness and/or availability of certain activities. This
level also includes one’s psychological orientations, which consist of three subjective
evaluations: 1) one’s beliefs about what s/he and the other ought to do, 2) what one likes
or wants to do, and 3) the extent to which an individual has the competence to perform a
certain behavior. Depending on these assessments, an individual may decide to not move
beyond this level of barriers. Crawford, Jackson, and Godbey (1991) viewed
intrapersonal constraints as the most powerful, because they control one’s will or
motivation to act. However, these barriers are also believed to be relatively unstable and
can be modified over time (Crawford & Godbey, 1987).

Interpersonal barriers involve the interaction between individuals or the
relationship between individuals’ characteristics. These barriers can arise from joint
preference for certain activities, or availability of a suitable partner. Structural barriers are

those constraints that act between activity preference and participation. These barriers

20

include financial resources, work schedule, availability and knowledge of opportunity,
and reference group attitudes regarding the appropriateness of the activity.

In this model, an individual does not encounter constraints at the interpersonal
level until any constraints at the intrapersonal level have been eliminated, and s/he does
not encounter structural barriers until there are no interpersonal barriers remaining.
Participation then occurs if there are no structural barriers, or if they have been
successfully negotiated. However, these barrier types are not independent of each other.
Some constraints may be found at different barrier levels, such as appropriateness and/or
availability of an activity (Crawford & Godbey, 1987). At the intrapersonal level, these
factors may act on an individual’s desire to participate in an activity (it is either
unavailable or perceived to be inappropriate), while at the structural level, an individual
may wish to participate but feel unable to do so because of inappropriateness or lack of
availability. In sum, the three levels are interdependent and overlap at times. The
sequential and overlapping nature of this model is different from models that have
previously been used to study DV and health care and it was the incorporation of these
qualities that makes this study’s Barriers Model unique.

Physician barriers to DV screening. A number of barriers to DV screening have

 

been identified through surveys of health care providers. A comprehensive list of such
barriers was provided in a review of the literature by Gremillion and Kanof (1996). They
divided barriers identified by physicians into four categories: contemporary social issues,
personal factors, professional factors, and institutional and legal factors. Social issues
include societal tolerance of domestic violence, desensitization through exposure to DV,

and prejudicial attitudes that DV is limited to families living in poverty. Person factors

21

may include a personal history of abuse, privacy concerns, and a sense of powerlessness.
Professional factors cited included time constraints, inadequate skills, and professional
detachment. Institutional and legal factors include limited institutional resources and a
potential loss of patient insurance if DV is reported. Health care professionals also cite
financial disincentives and marginalization by colleagues as barriers to working on DV
issues (Cohen, De Vos, & Newberger, 1997). Some of these barriers, such as lack of
education on DV, have been found to directly impact rate of screening (Parsons, Zaccaro,
Wells, & Stovall, 1995) and can be directly addressed through provider training.

Another often cited reason for hesitance to screen for abuse is fear of offending
the woman, particularly in private clinics with patients from higher SES backgrounds
(Stewart & Cecutti, 1993). However, evidence suggests that this fear is unfounded. In one
study, not only did women not appear to be offended, but several women (including older
and well-educated women) spontaneously wrote positive comments on the survey and
others expressed relief at receiving support and information (Stewart & Cecutti, 1993).

Attitudes of classism, racism, and sexism can be significant barriers to effective
screening and intervention. Some health care providers might be too accepting of
violence in minority communities, believing it is an accepted part of life for some women
of color. Physicians may not follow-up on suspected abuse because they believe these
women do not want or will not accept help (Cohen, De Vos, & Newberger, 1997).
Battered women have been greatly banned by the acceptance of culture as an excuse for
wife-beating among members of the legal and social service communities in this country

(Huisman, 1996).

22

Finally, some providers identify the patients’ unwillingness to disclose or leave
their abusive partners, and fi'ustration at not being able to change the situation, as barriers
to DV screening (Gerbert, Caspers, Bronstone, Moe, & Abercrombie, 1999; Rodriguez,
Bauer, McLoughlin, & Grumbach, 1999). In these situations, providers may not ask about
abuse even when it is suspected, because they anticipate the patient will deny it or it will
not have any positive impact. In one study, some physicians reported not screening
because they did not feel they could adequately address the problem if it was disclosed,
while others reported “burnout” from screening for DV not resulting in disclosures
(Gerbert et al., 1999).

Patient barriers to disclosure of DV. There are also barriers to DV screening and

 

disclosure that may exist from the patient’s perspective, and until recently, these barriers
were largely ignored. One potential barrier arises from the extreme feelings of shame that
women in abusive relationships often experience. This shame may lead women to conceal
signs of abuse from their physicians (Liebshutz & Rich, 1998; McCauley, Yurk, Jenckes,
& Ford, 1998; Rodriguez, Szkupinski-Quiroga, & Bauer, 1996). Women may also be
impacted by their desire to keep the family together (Rodriguez et al., 1996). In addition,
battered women may have a justifiable fear of repercussions fiom the abuser.

A major barrier to accessing medical care is the abuser’s preventing his partner
from seeking such care (McCauley et al., 1998). When considered in the context of
prenatal care, a woman’s knowledge that she is limited in her ability to seek care at all
may create a barrier to her disclosing the abuse to her provider. She may fear that if the
abuse is discovered, she will not be able to obtain any prenatal care, putting her unborn

child at risk.

23

Some women may also fear that the physician will contact the police or other
authorities. This fear may be more prevalent in areas where mandatory reporting by
health care professionals exists, such as in California (Rodriguez et al., 1996). Increased
trust in the provider may facilitate disclosure by reducing such barriers. Liebschutz &
Rich (1998) reported, “The quality and duration of the patient-physician relationship,
including both primary and episodic care, created a level of trust which influenced the
ease of disclosure” (p. 108). Women reported having increased trust and less fear of
disclosure when they believed their physician already suspected abuse and when the
physician affirmed that other women shared similar experiences. Similar sentiments have
been expressed specifically by Latina and Asian women (Rodriguez, Bauer, Flores-Ortiz,
& Szkupinski-Quiroga, 1998), illustrating that these issues are cross-cultural. By listening
to patients, and providing support and referrals, physicians can provide a supportive
atmosphere for battered women. Thus, developing a trusting relationship with patients
will remove many barriers to successful DV screening.

When women perceive physicians to be uncomfortable discussing DV, they may
be less comfortable with disclosure. In particular, Latina and Asian women expressed the
desire to have providers initiate the discussion and to show an understanding of DV
(Rodriguez et al., 1998). Focusing only on physiological symptoms and ignoring
psychological, social, and spiritual problems can prevent the development of a trusting
relationship. Furthermore, patients report the need for providers to have a compassionate,
supportive, and nonjudgmental attitude (Rodriguez et al., 1996). These factors are

necessary to build the trust needed to disclose abuse.

24

Perceived quality of health care is negatively impacted by domestic violence. In a
study by Plichta, Duncan, and Plichta (1996) only 9.7% of abused women had ever
discussed their abuse with a physician. women who had experienced abuse in the past
year were twice as likely to report that their physician did not spend enough time with
them or make an effort to have them explain their problems completely. They were also
significantly more likely to describe their physicians as less knowledgeable and
competent in treating their illnesses, not answering questions honestly or completely, and
not ensuring they understood the information about their medical condition. Overall,
battered women had significantly more trouble communicating with their doctors than
nonabused women did. Poor communication was related to dissatisfaction with care, a
finding that is supported by other studies which have found good patient-physician
communication, specifically empathy, sensitivity, and other qualities of openness,
strongly related to patient satisfaction (see Plichta, Duncan, & Plichta, 1996). When
battered women are blamed for their abuse, or ignored by their health care provider, they
see health care professionals as ineffective sources of help with regard to their violent
relationship (Langford, 1996). A battered woman may be told her medical complaints are
“in her head,” decreasing the likelihood that she will disclose her abuse.

Integrated Model of Barriers to Effective DV Screening. The Hierarchical Model

 

of Leisure Constraints can be adapted to understand barriers to DV screening by replacing
leisure participation with DV screening or disclosure. Thus, the final goal is the behavior,
with several potential barriers to actually completing that behavior. Using the same basic
model, and integrating concepts from the battered women’s barriers model and empirical

research on DV screening in health care settings, the model found in Figure 1 was

‘25

developed. This model can be used to understand barriers to providers conducting DV
screening as well as barriers to disclosure of DV by the patient. It provides a theoretical
framework for training providers on DV, for understanding the interpersonal process of
DV screening, and for creating a protocol in a medical setting that might require

procedural adaptations to occur in order to make DV screening effective.

Structural

lntemersonal

Intrapersonal X
Provider/ Patient ¢ Screening/
// Disclosure

Figure 1. Potential barriers to domestic violence screening and disclosure in the health

care setting.

The intrapersonal barriers for the patient include shame, knowledge of available
resources, and attitudes about relationships and DV. Providers may not wish to screen for
DV because of lack of knowledge, attitudes towards victims of DV, their own

experiences with violence, and their perceived skill at conducting screening. For both

26

patients and providers, culture can play a role in many of these intrapersonal factors,
affecting attitudes, values, beliefs, and behavior.

The intrapersonal barriers play a vital role in this integrated model. The role of
intrapersonal factors can be found in otherisocial science models as well, although
different labels are often used. For example, in a model adapted from Social Cognitive
Learning Theory, Gagan (1998) attempted to explain barriers to accurate diagnosis and
intervention with domestic violence. According to this model, behavior (diagnosis and
treatment) is impacted by cognition, other personal characteristics, and environment.
Many of the variables of interest in affecting behavior would fall into the intrapersonal
level in the integrated model introduced here. Behavior change will not occur without the
necessary personal factors established first. Although there is overlap between levels and
navigating the barriers is not always linear, it seems that once the barriers at the
intrapersonal level are addressed, it is easier to navigate barriers at the interpersonal and
structural levels.

The interpersonal barriers in the situation of DV screening focus primarily on the
interaction between patient and provider. A patient’s willingness to disclose abuse may
be affected by trust in the provider, support offered, and perception of the provider’s
attitudes. For the patient, the interpersonal level may also include fear of reprisal from the
batterer. Providers may hesitate to screen for DV for fear of offending the patient or
invading her privacy. The influence of colleagues can also be a factor, either in increasing
or decreasing the acceptability of engaging in screening behavior. There is also a power
differential between patients and their health care providers, which can affect interactions

depending on how the power is used or misused. A misuse of that power can actually

27

replicate the dynamics of an abusive relationship, which would greatly inhibit a patient’s
willingness or ability to disclose abuse to her provider. Culture is included again at this
level, for it can have a significant impact on an interaction, depending on cultural
similarities or differences. Cultural differences between patient and provider can increase
the power differential, inhibit the development of trust, and affect perceptions of the
other.

Finally, the structural level encompasses many of the professional barriers that
have been identified in the empirical literature, such as time constraints, limited resources
for abused patients, marginalization by colleagues, and institutional policies and response
regarding DV issues. In total, the overriding atmosphere and activities of the health care
setting can set up barriers to effective DV screening. Culture comes into play again at this
level, in the way that the overriding culture of the setting can influence individual
attitudes and behavior. In this situation, culture refers less to factors such as individuals’
ethnicity and more to issues such as the general culture of the health care field, in contrast
to non-medical settings. F urtherrnore, racism, classism, sexism, homophobia, and other
forms of discrimination may be present within the system of the clinic, potentially
affecting the screening process.

All potential barriers should be considered when attempting to institute a DV
screening program in a health care setting. The implementation of the screening program
will likely proceed with fewer complications if the barriers are eliminated prior to
implementation, and if providers’ and patients’ concerns are addressed in the design of
the program and in the training itself. Several barriers, such as lack of training and

referral sources, can be remedied through a collaborative relationship with other local

28

agencies. Providing brief, yet informative, training on DV issues and screening
procedures, along with providing referral information, can address potential barriers at all
levels. Training can increase knowledge and skills, as well as change attitudes about DV,
to facilitate screening. These changes will also impact the interpersonal interaction
between provider and patient, since the provider will potentially be more knowledgeable
and empathetic towards an abused patient. Through training, providers may also realize
that DV screening does not require as much time as they believed, and once a screening
program has been implemented in a clinic setting, the overall atmosphere may become
more accepting and supportive of addressing DV issues. The long-term goal of such a
program is often to change the culture at the structural level to be more focused on the
prevention and treatment of domestic violence.

Culture can be seen to act at all levels of barriers to DV screening. Almost all
aspects of the intrapersonal are either part of, or impacted by, one’s culture. Some
significant barriers to DV screening and reporting may also be related to cultural issues
acting at the interpersonal level, and miscommunication or misunderstandings in the
context of discussion about DV experiences. Awareness of the potential for cultural
differences and a willingness to listen to alternative perspectives can minimize their
potential negative impact on the provider-patient relationship, specifically in the context
of DV screening. These issues should be of particular consideration if providing health
care to a diverse patient population. Once these issues are successfully addressed, the

many influences of culture at the structural level can be of greater focus.

29

Cultural Issues and Domestic Violence

 

Violence against wives occurs in more societies than any other form of family
violence, and is more common in societies where women have inferior status (Reichert,
1991). While DV occurs in almost every culture worldwide (Fischbach & Herbert, 1997),
and women of different racial/ethnic groups in the US. are equally likely to be victimized
by an intimate partner (Bachman & Saltzman, 1995), there are issues specific to different
cultural groups that professionals working with these women should be made aware.
Health care providers should be knowledgeable concerning issues that their patients
experience so that they might provide the optimal treatment, as well as referring women
who report abuse to additional communityresources. Unfortunately, many lower-income,
rural, inner city, minority, and immigrant communities in the US. are lacking in
culturally appropriate services for the prevention and treatment of domestic violence
(Pinn & Chunko, 1997).

Training on DV issues should include a discussion of cultural issues. King et a1.
(1993) explained that cultural factors extend beyond race or ethnicity to include economic
situation, family structure, gender roles, marriage patterns, sexual behavior, contraceptive
patterns, pregnancy and childbirth practices, child rearing practices, housing
arrangements, migrant status, occupations, use of chemical comforters, self-treatment
strategies, and lay therapies. All of these factors together define how experience is
organized; specifically, what each woman’s subjective experience of being abused will be
(King et al., 1993). While becoming educated on some cultural issues is an important step
for a health care provider to take, it is also important to assess each individual’s degree of

acculturation into the dominant culture and degree of identification to her own culture.

30

It is important to recognize how the constraints of racism, sexism, classism, and
other forms of oppression might impact a woman’s ability to cope with an abusive
relationship (Phillips, 1998). Not only might a woman’s minority racial/ethnic status limit
her resources, but it is necessary to also consider her language skills, economic situation,
and religious experience in determining the amount of power and resources available to
her. For example, an African American woman living with her abuser and receiving
social assistance will have different resources available to her than a middle-class African
American woman who owns her own home.

The barriers to effective detection of DV in the health care setting that were
discussed previously may be heightened when working with poor or minority women
(Pinn & Chunko, 1997). Additional barriers may include unavailability of a translator,
traditional beliefs accepting violence as the man’s choice, increased shame, or fear of
consequences from immigration or governmental authorities.

There is a high probability that the provider will not be of the same ethnic
background as the patient, due to the relatively small number of minority health care
professionals. One barrier that may be common to all women of color is that her cultural
group may be further stigmatized by her disclosure of abuse to a health care professional
(Campbell & Campbell, 1996). For example, an African American woman may not wish
to reinforce the stereotype that African American men are violent (Phillips, 1998). The
woman may also be concerned that the provider will call the police. This concern stems
from a history of discrimination against men of color by the criminal and judicial systems
and often prevents many women from seeking assistance (Campbell & Campbell, 1996).

Finally, in the context of serious health problems and economic survival, DV may not

-31

seem to be of the utmost importance to an abused woman of color. Women of minority
groups are more likely to be poor, face discrimination in employment and housing, and
have less education than white women (Phillips, 1998). Poverty increases the obstacles
for abused women, as well as their vulnerability for abuse, and often maintains situations
of abuse (Phillips, 1998). These factors resulting from poverty may outweigh the
importance of DV for these women, and may affect the likelihood that an abused woman
will seek DV intervention services. With a disproportionate number of minority women
living below the poverty line, these issues are valid, and should be respected by the health
care provider. However, a woman’s decision to not seek professional services should not
preclude health care providers from continuing to discuss DV with her.

Health care professionals should be aware that domestic violence interventions
and other community resources may not be sensitive to cultural diversity (Heron, Jacobs,
Twomey, & Kaslow, 1997), and a patient’s reluctance to seek services may be from
previous experiences with such insensitivity. Service providers are commonly seen as
more tolerant of DV by or against people of color, as well as displaying little respect
towards women of minority ethnic groups (Sorenson, 1996). Not only do women of
minority ethnic backgrounds have to cope with the traditional sexism of their culture, but
they must also face sexism and racism in the US. Some women may sacrifice sexual
equality in order to avoid racism outside of their community (Huisman, 1996). Rather
than blaming a woman who does not follow through with referrals, the health care
provider can try to assist the woman in finding culturally appropriate services.

Domestic violence does maintain many similarities cross-culturally. Sorenson

(1996) found that a central theme in studying violence against women across cultures is

32

the influence of social and economic inequalities on interpersonal relationships. Across
cultures, the most significant predictor of DV is this power differential in relationships
that results from such inequalities between men and women (Crites, 1990). Furthermore,
the pattern of multiple forms of abuse, including physical, psychological, and sexual
abuse, is similar across ethnic groups (Sorenson, 1996).

In cultures which value the family or kinship over the individual, the extreme
shame commonly felt by abused women may be intensified, and may lead to mental or
spiritual health problems (Fischbach & Herbert, 1997). These problems can be
manifested in different ways, but may ultimately lead to the same outcome. For example,
one out of four white American women who attempt suicide have been abused, as have
half of all African American women who attempt suicide (Fischbach & Herbert, 1997). In
cultures where the shame of DV is even greater than in the dominant culture, suicide may
be seen as the only alternative to abuse.

The following discussion will focus on four groups of minority women: African
American, Hispanic, Asian and Asian American, and immigrants. While there are
numerous other groups of women in the US. that are also victims of intimate partner
violence, these groups are of focus because they constitute the majority of minority
patients seen in the clinic in which this project was implemented. Prevalence rates of
abuse are similar across ethnic groups, but there is some evidence that certain sub-groups
may be at greater risk for DV than others. This issue will be discussed in greater depth in
the following sections.

African American Women. There are several theories explaining why there is

 

violence within African American families. One theory examines the conflict between

33

African and American gender roles (Ucko, 1994). There is greater flexibility .between
gender roles in African American families, with women often assuming the role of
breadwinner and men often sharing in housekeeping. These dynamics are an extension of
African traditions, as well as necessity during slavery. Traditionally, African women
received more respect and held more power in society than women in modern cultures,
and male violence against women was not prevalent. These values were in direct conflict
of those European values to which Africans were introduced upon becoming enslaved,
and the conflict has continued since. Today, black couples are more egalitarian than white
couples, although there is now evidence of wife abuse in African American families.
While the European influence created a greater power imbalance between black men and
women, African men were not given the status or power of white men, creating a lasting
conflict and forcing African American men to seek alternative sources of power and self-
worth. It has been proposed that African American men are violent towards women
because of their lack of power within the dominant culture and related loss of self-esteem
(McGee, 1997). Furthermore, the introduction of the European view of male dominance
in the home produced conflict within marital relationships that had traditionally been
based on equality.

African American women may be less likely than white women to disclose DV to
their health care provider, because of pressures to not share family problems with
outsiders (Norton & Schauer, 1997). There is some evidence that African Anierican
women are more likely to disclose DV to an African American health care provider
(Heron et al., 1997). Because of the history of institutional oppression that they have

experienced, there is often a desire to keep problems to oneself, or at least within the

34

community. There may also be a fear of reinforcing negative and racist stereotypes by
revealing violence within the community (Campbell & Campbell, 1996). Black battered
women are less likely to use shelter services than white women, possibly because of the
perception that shelters are for white women and that the battered women’s movement
represented the interests of white women (Coley & Beckett, 198 8). Local shelters often
do attend to the needs of black women, but some women may be unaware of the available
services. In this case, the health care provider can be a valuable source of information for
the abused patient.

Lower SES African American women report more marital violence than middle
and upper class black women (Lockhart & White, 1989). For lower-income black
women, economic survival may take precedence over seeking DV services. Therefore, the
health care professional should be prepared to provide referrals not only for shelter or
counseling, but also for such services as legal aid, childcare, emergency housing, and
educational and employment opportunities (Coley & Beckett, 1988). The health care
professional should also recognize that leaving the abusive relationship is often not the
initial goal for a low-income African American woman, and imposing this priority on her
may be contradictory to her cultural values (Heron et al., 1997).

African Americans generally have an extended kinship system that is their
primary social support system. There is some evidence that the higher the amount of
family and community involvement, the lower the amount of DV (U zzell & Peebles-
Wilkins, 1989). Lower SES black women, in particular, more often turn to friends or
relatives than to formal resources for help when they are abused (Lockhart & White,

1989), making disclosure to a health care provider quite meaningful. It may be that a

35

woman does not have a social support system available, or she may be experiencing such
severe violence that she decides to seek help outside the community. Some of these issues
are also relevant to women in other minority communities, including Hispanic and Asian
American women.

Hispanic Women. There is great variety within the Latino culture, along with

 

some similarities across subgroups. Many studies examining DV across ethnic groups
have failed to consider differences within the Hispanic culture or level of acculturation
(Kantor, Jasinski, & Aldarondo, 1994). The three primary Hispanic subgroups in the
United States are Puerto Rican, Mexican, and Cuban. There are significant differences
between these subgroups in acculturation, impoverishment, and acceptance of wife abuse
as normative. In addition, it has been found that Mexican and Puerto Rican husbands
born in the US. perpetrate more DV than those men born in the native country. Among
Cuban families surveyed, DV was practically nonexistent, according to their report
(Kantor, Jasinski, & Aldarondo, 1994). Sorenson and Telles (1991) found evidence that
Mexican Americans born in the US. experienced significantly more DV than those born
in Mexico and non-Hispanic whites born in the US. They proposed that the conflicting
demands from their culture of origin and the dominant culture may lead to this increased
rate of violence.

Latino culture values family and group achievement over individualism (McGee,
1997). The family orientation is also characterized by an extended kinship system.
Traditionally, gender roles are rigid and the power differential within families is large.
Traditional Mexican American, Cuban, and Puerto Rican cultures subscribe to the idea of

male dominance and female obedience within families; ideas which are associated with

36

violence against women (Sorenson, 1996). Although the traditional family values
discourage DV, these power dynamics have resulted in abuse in the Latino American
community (McGee, 1997). It has been proposed that Mexican American women
maintain a more tolerant attitude toward DV than Anglo-American women because of a
cultural tradition giving men control within the family (Olavarrieta & Sotelo, 1996).
Some Mexican Americans have described their culture’s reverence of men’s sexual
prowess, in contrast to the emphasis on a woman’s virginity, in defining gender roles
(Sorenson, 1996). In addition, several factors associated with increased risk for marital
violence, including low socioeconomic status, unemployment, and income disparity
between husband and wife, are more common in Hispanic-American families than non-
Hispanic families (Kantor, Jasinski, & Aldarondo, 1994).

Mexican American women born in Mexico may face a greater risk of marital rape
than non-Hispanic white women (Sorenson & Telles, 1991). Of Mexican-born Mexican
American women reporting a history of adult sexual assault, 31% were assaulted by a
husband or partner. In contrast, the same circumstances were reported by 13% of non-
Hispanic white women. Additionally, high rates of emotional abuse (80%) were reported
by one sample of Latina (predominantly Mexican) women (Santiago & Morash, 1995).

The additional problems faced by Mexican families who have immigrated to the
US. may also lead to increased risk for DV. Mexican women may be isolated because of
a language barrier, they often lack sufficient education to obtain well-paying jobs, and
their fear of deportation may prevent them from seeking help when they are abused
(Olavarrieta & Sotelo, 1996). In Mexico, low levels of education and large families

(seven or more children) are correlated with domestic violence (Olavarrieta & Sotelo,

37

1996). In addition, women living in rural areas with low socioeconomic status are more
vulnerable to physical abuse. Families in rural areas tend to hold to traditional gender
roles to an even greater extent than Mexican culture as a whole. There is a reported DV
prevalence rate of 33% in Mexico, with similar rates of abuse during pregnancy
compared to US. statistics. Because of outdated laws viewing DV as a private matter, as
well as other numerous legal barriers, it is believed that many cases go unreported. The
reported rates are believed to be similar in other Latin American countries, some of which
have made great progress in criminalizing DV and protecting women’s safety.

Before referring women to community agencies, it would be useful to know what
interventions are effective. It is estimated that only 2% of all battered women enter the
shelter system, and anywhere from 7% to 56% of DV is reported to the police (Wiist &
McFarlane, 1998). A small number of Latina women experiencing emotional or physical
abuse indicated that they would seek “formal” help for problems with their partners,
while most (73%) would seek help from family members (Santiago & Morash, 1995). In
reality, many women do not tell anyone about the abuse, and the help they do receive
from family members may not be sufficient to stop the abuse. Some women will turn to
social service agencies for help, particularly if they are specific to the Hispanic
community.

It is logical to think that a woman’s race or ethnicity might influence the services
she seeks, as well as considering the influence of pregnancy on help-seeking behaviors. In
one study, pregnant Hispanic women who had used the police experienced more severe
abuse than those who did not use the police (Wiist & McFarlane, 1998). Those women

who were older, with higher incomes, and severely threatened with abuse used the police

38

most frequently. Just over half (51%) of the women that did use the police reported they
were very effective in helping to reduce the violence, while 17% reported that reporting
to the police had made the violence worse. Violence did end in more families where
police involvement had occurred than in families where the police were not used, but
increased violence is a potential consequence of police reporting that must be considered.
In addition, few services other than police were utilized by battered pregnant Hispanic
women, and discovering why these women are not seeking commrmity resources is
important before making community referrals.

According to Santiago and Morash. (1995), lower use of community resources by
battered Latina women may be attributed to cultural norms emphasizing help-seeking
within family networks, community failure to recognize diversity in the Latino
population, limited access to services, and limited legal protection for immigrants. These
factors may inhibit a Latina patient’s comfort in disclosing DV to a health care provider,
but conveying sensitivity to cultural issues will increase the chances of disclosure.
Similar barriers to professional help-seeking may be found in other minority women as
well.

Asian and Asian-Atrrerican Women. For summary purposes, Asian culture will be
discussed broadly. However, it is important to remember the diversity within Asian
subgroups. Furthermore, this discussion will include both Asians and Asian Americans,
because of the overlap in the research literature. The population of interest in the current
study includes both Asian American women and Asian women who are temporarily in
the United States for educational purposes, so issues of both groups are relevant to this

project.

39

Within the traditional Confucian social system, Asian families were patriarchal in
their power structure, with men being valued more highly than women (McGee, 1997).
The traditional values also expect individuals to put the good of the family ahead of their
own desires or needs, with the family seen as more important than the individual. It has
been proposed that overall, domestic violence is present in the Asian American
community because of the power differential between men and women, the lack of access
to resources for women, the pressure to maintain harmony within the marriage, and
sexism within the culture (McGee, 1997). Violence in an Asian family may also result
from a woman taking advantage of the increased opportunities for freedom offered in the
US. (Crites, 1990). Many cross-cultural marriages between US. servicemen and Asian
women have occurred, and the conflict that may arise upon return to the US. has also
been associated with violence.

There is evidence that Asian women are less likely than women fi'omother ethnic
groups to report abuse, and it is often severe by the time they do (Huisman, 1996). This
tendency may be tied to traditional Asian values, including patriarchal attitudes. Shame
plays a strong role in Asian cultures, which may cause Asian women to experience
increased psychological damage following abuse (Crites, 1990). Not only would
unacceptable behavior bring shame on oneself, but upon the entire family and ancestry as
well. Suicide has been found to be much more prevalent in abused Asian women,
possibly because of the effects of this tradition. Asian women may also not disclose
experiences of abuse because of fear of being ostracized from their communities.

Korean immigrant families appear to be at particularly high risk for DV, with

reported prevalence rates almost double national U.S. statistics (Rhee, 1997). Although

40

many Korean immigrants have higher levels of education and come from higher
socioeconomic status than many other immigrants, most first-generation Korean
immigrants do not learn English and face limited job opportunities. Although many
Korean women are in the work force full time, many families still adhere to traditional,
rigid gender roles of male dominance and female submission. Surprisingly, divorce is
common among Korean immigrants, with physical violence by the husband cited as the
leading cause.

Social services have been found to neglect the needs of the battered Asian woman
(Huisman, 1996). Language is a barrier particularly difficult to overcome in this
community, due to the variety of Asian languages and dialects that are represented. Some
women may be hesitant to seek any services because of their difficulty in commrmicating
with people outside their community. Even women who speak fluent English may prefer
to work with someone from their ethnic group. Different outreach and advocacy activities
may be necessary to i'each Asian women and to make them feel comfortable seeking
services.

As with other ethnic groups, number of years and generations in this country, and
level of acculturation should be taken into account when considering the influence of
culture on an individual’s life. For example, in a study of Chinese Americans’ attitudes
toward DV, age was an important factor in determining tolerance for violence, with
younger respondents being less tolerant of partner violence (Yick, Agbayani-Siewert,
1997). These younger participants may have abandoned some of the more traditional,

patriarchal, attitudes of Asian culture.

41

Asian Indian culture also traditionally devalues women and places men in the
dominant role in families, contributing to violence against women in India (Dasgupta &
Warrier, 1996). Furthermore, despite the fact that the Indian community in the US. is
largely educated, skilled, and financially stable, DV continues to be a significant problem
(Dasgupta & Warrier, 1996). Although Indian families tend to value education highly,
even for their daughters, there remains an emphasis on finding a suitable husband and
making a good home. In some cases, a good education is simply seen as a means of
finding a good husband. Dowry continues to be granted to the husband’s family, and
dissatisfaction with the dowry is ofien an excuse for abuse. Dowry-related death
continues to occur with some frequency in India (Fischbach & Herbert, 1997). Although
dowry-related homicide may not occur in the US, the underlying dynamics of DV in the
different culturescan be seen as parallel.

In addition, although many Indian women are professionals with well-paid jobs,
there is a tendency for the husband to control the financial resources, and for significant
financial abuse to result. Physical abuse did not stop during pregnancy in one group of
abused Indian women; for many, the abuse increased and was directed at the womb
(Dasgupta & Warrier, 1996). For some Indian women, sacrificing their own well-being
for the good of the family, especially the children, is common. Fear of being ostracized is
ofien a common reason for not disclosing abuse in this community.

Immigrant Women. Women who are not yet citizens and who do not speak

 

English fluently are at an increased risk for intimate partner violence (Phillips, 1998). For
instance, a batterer may use his wife’s immigration status as a way of intimidating and

exerting power over her (Pinn & Chunko, 1997). Many immigrant women are forced to

42

stay with abusive partners, despite changes in immigration laws in the 1990sthat
attempted to address this problem. The passing of the Violence Against Women Act

(V AWA) in 1994 was also expected to make it easier for immigrant women who are
victims of DV to seek assistance without fear of deportation, but more research needs to
be done to evaluate this relatively new policy. In addition, although laws in the US. may
have recently changed, many women are unaware of these changes and their access to
such information may be controlled by an abusive partner.

Immigrant women are often poor, uneducated, may not speak English, and do not
yet have legal residency, placing \many constraints on their ability to leave an abusive
relationship (Phillips, 1998). These women often depend on their husbands for
communication and information concerning community resources. F urtherrnore, an
immigrant woman may come from a traditional culture which does not support
involvement of outsiders in private matters. Leaving their country of origin also brings a
loss of ties to extended family and friends, increasing isolation (Sorenson, 1996). In
addition to being isolated, the stress of immigrating and adjusting to a new life can
increase the risk for DV (Jang, Lee, & Morello-Frosch, 1990)

In situations where patients speak a language or dialect unknown to the provider,
attempts should be made to find an interpreter other than a family member (Campbell &
Campbell, 1996). Although the family member may be easiest, it is not recommended
when dealing with issues of violence. For professionals working with immigrant women
who are being abused, both these particular issues, as well as the issues discussed above
for women from specific ethnic backgrounds, need to be considered. These women are

less likely to know of their rights and may need additional guidance in getting help.

43

Although the rate and types of violence perpetrated against women may be similar
across cultures, there are differences in how the violence is interpreted, dealt with, and
treated among different ethnic groups. As stated by Ganley (1998):

Victims from different cultural groups have different values and beliefs

about interpersonal communication, the role of health care providers, .the

role of police, and the role of family members which shape how they

reveal or don’t reveal their experience of domestic violence (p. 27).

It is important for professionals working with battered women to recognize the potential
for these differences, in determining how to best identify and help a woman who is being
abused. This project was designed to assist clinic staff in recognizing such differences
and feeling skilled in working with pregnant women of all backgrounds who may be
experiencing domestic violence. As stated, cultural issues can serve as barriers to DV
screening/disclosure at all levels. It was hoped that increased awareness of this potential
would decrease the impact of these cultural barriers.

Conclusions

The need for routine domestic violence screening of pregnant women is evident.
One question that remains is how culture may act as a barrier to DV screening, and if a
universal measure and technique is effective cross-culturally, or if different procedures
need to be developed when working with women from different cultures. Women usually
have numerous and regular contacts with their obstetricians over the course of a
pregnancy, providing the ideal opportunity for doctors to discuss the possibility of abuse

with their patients. Many abused women report they would not volunteer information

about DV to their physician, but they would answer honestly if asked directly. It appears

44

that asking women about abuse directly and repeatedly over the course of pregnancy
increases the chance of disclosure. While it is not the doctor’s role to counsel an abused
woman, all health care providers should be given information on how to discuss DV_with
their patients and where to refer women for additional intervention should a woman
choose to seek additional services. Furthermore, they should be aware of the potential
barriers that may prevent a woman from disclosing abuse. However, the basic
information provided in medical training programs often does not include a discussion of
DV against minority women, and there are specific issues that health care providers
should consider when discussing DV with patients from minority groups, as well as what
resources exist that are appropriate for each individual.

While many programs across the country have developed DV screening protocols,
there is little information in the research literature on evaluation of these programs. Most
research consists of establishing prevalence rates based on a certain screening tool. This
project adds to health care services research by providing additional evaluation
information, as well as presenting ideas for improving services. Although the primary
focus of the evaluation was the expanded screening program and issues surrounding
cultural differences, the clinic also had their current procedure evaluated, in terms of
current disclosure rates and provider knowledge, attitudes, and behavior.

In addition to the valuable information provided by the evaluation of the tmiversal
screening program, this study also contributes to the literature by examining the
theoretical underpinnings of the DV screening process. First, most studies of DV
screening programs do not incorporate a theoretical model to understand the dynamics of

screening and disclosure. The model representing barriers to screening and disclosure

45

provides a rationale for the training program and for the hypotheses concerning patient
disclosure of DV. Second, this study helps with understanding the differences between
women of different racial/ethnic groups in their level of comfort with DV screening and
willingness to disclose abuse experiences. While some research has focused on DV in
specific ethnic minority communities, this study takes a different approach in examining
differences within a group of pregnant women, through a combination of self-report
survey data and rates of screening and disclosure, from both the dominant culture and
minority groups. The Barriers Model also aids in the interpretation of these data.

The overall goal of this project was to identify and decrease the impact of barriers
to DV screening and disclosure, based on a theoretical barriers model, to achieve a
successful DV screening program. The project was intended to directly affect the health
care providers, but not the patients. It was believed that patients’ comfort in disclosing
DV would be indirectly affected by changes in the providers’ behaviors. The objectives
were to 1) train ob/gyn staff on DV screening, thereby decreasing barriers to screening; 2)
increase screening and disclosure of DV; 3) assess the role of race/ethnicity as a barrier in
the screening process; 4) identify patients’ views on DV screening and their perceived
barriers to disclosure; 5) assess the impact of training on providers’ knowledge and
attitudes; and 6) provide intervention for abused women. The research hypotheses are

based on the project goals and objectives.

46

HYPOTHESES

. Following provider education and training, patients will be screened for DV
throughout their pregnancies. This procedure will lead to a significant increase in the
number of times women are screened and the rate of disclosure of violence.
Exploratory analyses will determine if differences exist in overall screening or
disclosure rates across ethnic groups, to determine if race/ethnicity serves as a barrier

to effective DV screening.

. Women of color who disclose abuse will do so at later visits than Caucasian women.

Women of color will require more time to develop trust in their providers, while both
Caucasian women and women of color will be equally likely to experience late onset
of abuse.

. Providers’ race/ethnicity will affect overall patient disclosure rate and time of
disclosure. Women will be more likely to disclose abuse to a provider of the same
race/ethnicity as herself. Exploratory analyses will determine if provider gender
impacts patient disclosure.

. Overall, patients will support the concept of universal DV screening and report
feeling comfortable with the process, but women from minority groups will report
more barriers to disclosure and less overall comfort in disclosing abuse to a health
care provider.

. Domestic violence education will reduce barriers to conducting DV screening by
increasing providers’ knowledge of DV and awareness of cultural issues related to

DV, changing attitudes towards DV and its victims, decreasing the number and

47

impact of perceived barriers to conducting universal DV screening, and increasing the
amount of self-reported DV screening as measured six months after training.
When women disclose abuse to their physicians, they will be referred to a social

worker for further intervention and community referrals.

48

METHOD

Because of the importance of the setting in implementing this project, it will be
described first. The procedure sections describe the methods chronologically, with
descriptions of participants included within each of the procedural sections. Finally, a
description of measures concludes the method section.

Settin

The Obstetrics and Gynecology Clinic at Sparrow Hospital is a training site for
twelve residents, each of whom complete a four-year residency program. There is a long-
standing nursing staff of eight, and two on-site social workers. The clinic primarily serves
low-income patients receiving Medicaid benefits. In 1999, 488 patients were seen for
prenatal care, with a total of 359 live births. This clinic was chosen as the project site
because it is a training clinic, which provides opportunities for educational interventions
that other health care settings may not. In addition, there is substantial racial/ethnic
diversity in the clinic’s patient population, allowing for the investigation of the impact of
this factor on DV screening. The hospital had ceased collecting race/ethnicity information
as part of the standard intake form and, consequently, these data were frequently
unavailable (or very difficult to find) in the medical charts. Therefore, race/ethnicity was
added to the abuse screening form for the purpose of this research project.

Prior to implementation of this project, the clinic nursing staff was already
conducting DV screening and intervention. According to clinic protocol, at the initial
prenatal visit women completed a written questionnaire containing questions'about abuse

history, which were then discussed verbally with either a social worker or a nurse. The

49

residents did not discuss DV with a patient unless she had previously disclosed abuse to a
nurse or social worker and it was indicated in the chart notes. If a woman did not disclose
abuse at the initial visit, or if she was not screened for any reason at that visit, she was not
screened again.

The clinic staff was receiving training on basic DV issues in the health care
setting as required for accreditation, but it did not address cultural issues related to DV or
effective screening for DV. Furthermore, the training had not been evaluated in any way
in this clinic, so its impact was not known.

Procedures

A timeline of project activities is presented in Table 1.

50

35362

.95

82823
bEzEEoU

“avenge—gun

 

geese eyes ._ 29¢

51

Preliminary Activities. In the formative stages of this project, the principal
investigator met with the residency program director, clinic manager, and clinic social
workers to determine if the proposed project would be useful and feasible. The residency
director agreed to the project and requested that the provider training go beyond basic DV
issues, to focus on multicultural issues and DV. It was also agreed that provider training
would occur during the residents’ regular training rounds, maximizing attendance.
Meetings with the clinic manager and social workers provided a better understanding of
clinic policies and procedures, and they contributed their perspectives on the provider
training and universal DV screening. Although they believed that the nursing staff did an
effective job at working with patients experiencing violence, they were pleased that the
residents would assume a more active role in the identification of abused women.
Furthermore, the nurse manager agreed to make attendance at the training workshop
mandatory for the nursing staff.

Before the project officially began, the residency program director resigned and a
new residency program staff was appointed. The new director was unaware of the
agreements that had been made concerning this project, and new discussions had to occur
in order to ensure successful continuation of the project. Overall, the director and
assistant director were supportive of the project goals, although some minor changes were
requested. A

To assist with data collection and other project activities, four undergraduate
research assistants were recruited. Their participation involved a three-semester
commitment and they were required to attend weekly project meetings. These meetings

provided time for training the research assistants on data collection methods and

52

education about different issues related to DV. They gained knowledge of all issues
pertinent to the project, including basic DV information, cultural issues, DV in the
healthcare setting, and basic research methodology. Furthermore, each research assistant
picked a topic of interest to her/him in the area of violence against women and led a
discussion on that topic. Their final project was to write a research proposal about that
topic.

Training of the research assistants focused on collecting data from the medical
charts and patient surveys. In addition to showing competence in collecting data, they
were required to demonstrate professionalism and sensitivity to the patients prior to
beginning data collection. Approximately one month of training was completed prior to
initiating data collection activities, and additional training occurred throughout the three
semesters.

Archival Data Collection. Prior to provider training, clinic procedure dictated that

 

all prenatal patients were screened for domestic violence through both written and verbal
methods, by a nurse and a social worker. Patients completed a written screening form that
had one question addressing physical abuse and one question addressing sexual abuse. At
the first prenatal visit, the nurse reviewed this completed form and followed-up on any
positive responses. There was also a form in the chart that the nurse completed after
verbally asking each patient about physical abuse. Finally, patients who were part of the
Maternal Support Services Program (MSSP) met with a social worker, who also assessed
for domestic violence. [Notez any woman receiving Medicaid benefits was included in the

MSSP.]

53

The first step in data collection was to establish a base rate of DV screening in this
patient population. Of the approximately 500 women seen for prenatal care in 1998, 300
were selected using a list of random numbers generated by Microsoft Excel. The 1998
patients were used to ensure that all had either given birth or terminated the pregnancy
prior to the onset of data collection. These 300 charts were reviewed for demographic
information and documentation of DV screening or voluntary disclosure, based on clinic
forms for nurse and social worker screening at the first prenatal visit.

The principal investigator and four research assistants collected 14 data items
from each medical chart. In order to ensure reliable data collection, the principal
investigator randomly selected at least 25% of the charts reviewed by each assistant and
found an overall agreement rate of 90%, with at least 80% agreement on individual
variables. Patient ethnicity was the only variable not reaching at least 80% agreement at
the first reliability check, due to the difficulty in finding this information in the charts (as
mentioned, race/ethnicity is not included on any of the hospital’s standard medical
forms). The discrepancies in ethnicity data were due to the principal investigator
identifying patient race more often than the research assistants did. The research
assistants had been shown several locations in the charts where race was sometimes
recorded, but were also instructed to not spend an unreasonable amount of time searching
for this information, as it was frequently not in the charts at all.

The second chart review occurred six months after the new screening procedures
began. Medical charts of all women seen for prenatal visits during those six months, who
had either given birth, terminated a pregnancy, or transferred care to another provider,

were reviewed to determine the rates of DV screening and disclosure. Similar to the

54

initial chart review, the second review was completed by the principal investigator and
four research assistants. The information collected from the charts was slightly different
from the initial chart review, so additional reliability checks were completed. When
inadequate agreement was found for two items (primary doctor and number of visits),
additional training occurred and the charts were re-reviewed. Following this procedure,
agreement reached at least 80% on each item and over 90% for the overall charts. It is
worth noting that the discrepancies in data’collection were primarily due to the method of
record keeping of the clinic and the fact that patients often see two or three different
doctors throughout their pregnancy, making identification of a primary doctor difficult at
times.

A total of 572 charts were reviewed: 300 charts prior to the staff training and 272
charts 6-8 months after the training. Ten of the charts in the second chart review were
later eliminated because patients had not had a prenatal appointment with a doctor prior
to giving birth. In the first chart review, race/ethnicity was only identified in 38.4% of
patients. Of those for whom race/ethnicity was known, 57.0% were Caucasian, 25.4%
African American, 9.6% Hispanic, 2.6% Asian/Pacific Islander, and 5.4% “other”
(includes Native American, Arabic/Middle Eastern, and Turkish). The mean age of
patients was 23.81 years (S_D. = 5.32).

In the second chart review, race was known for 75.6% of patients. Two factors led
to the increase in racial data: first, greater familiarity with the medical charts allowed for
faster and more accurate chart review and second, race/ethnicity was included on the
abuse screening form and was often completed by those physicians who did screen for

DV. Of those patients for whom race/ethnicity was identified, 61.1% of patients were

55

Caucasian, 20.2% African American, 10.1% Hispanic, 4.0% Asian/Pacific Islander, and
4.5% “other” (includes biracial, multiracial, Middle Eastern, and Native American). The
one Native American and two Middle Eastern patients were combined with the “others”
in order to increase cell sizes for data analysis. The mean age of this group of patients
was 23.91 years @ = 5.33).

The abuse data collected at the second chart review came from the Abuse
Assessment Screen that had been introduced at the provider training. Beginning the day
of the training, the front desk staff was instructed to insert the AAS in each pregnant
woman’s chart prior to her next appointment. Additional forms were placed in the
physicians’ work area so they would be able to easily retrieve one if a patient had been
missed by the front desk.

Provider Training. The training workshop was required for all clinic nurses and

 

residents. All clinic staff members, with the exception of one medical assistant, attended
the training. Prior to the DV training session, all clinic staff were asked to complete a
survey. Providers were informed that although completion of the surveys was not
required by the clinic, it would assist in improving health care services as well as
implementing the program as smoothly as possible. Some surveys were returned to the
researcher by mail prior to the training session and others were completed just before the
training and returned in person. Approximately six months after the training seminar,
providers were asked to complete post-training surveys. Changes in knowledge, attitudes,
perceived barriers, and screening behavior served as indicators of training impact, and

were indications of a reduction of intrapersonal barriers to screening for DV.

56

The training was conducted by the researcher and the community educator from
the local DV shelter. Much of the information in the training was adapted from the

Family Violence Prevention Fund’s (The Fund) Improving the Health Care Response to

 

Domestic Violence: A Trainer’s Manual for Health Care Providers (Ganley, 1998). The

 

Fund manual integrates information about DV with information about the impact of the
cultural context on health care. One of the underlying assumptions of the Trainer’s
MM is that barriers to effective responses to DV victims often involve a lack of
knowledge or skills by providers, as well as system policies and procedures. The
trainings, therefore, are intended to provide information to decrease the impact of these
barriers. There are five 90-minute modules provided in the Trainer's Manual that are
related but can also each stand alone. These modules are: 1) Domestic Violence: A
Primary Health Issue; 2) Domestic Violence: Cultural Competency; 3) Screening,
Assessment, Intervention, and Documentation; 4) Domestic Violence: Practical
Applications; and 5) Legal Issues and Community Resources. Each module combines
didactic and interactive learning techniques, and they are intended to be adapted to the
needs and resources of the setting. Ganley asserted that provider training must be
accompanied by institutional support, such as policies and procedures for intervening
with DV victims. The clinic in this project did have that institutional support, including
hospital policies and administrative support, so the focus of the training was on individual
provider barriers.

Because of limited time, the full-day Fund training had to be condensed. Rather
than focus on only one module, several modules were combined and abridged to fit into

the 90-minute training session. The community educator introduced the general DV

57

information and the researcher presented the information on health care issues and
screening techniques. The 15-hour training briefly reviewed general information about
DV, effects on women’s health and pregnancy outcome, myths, and overcoming barriers
to screening. An ongoing discussion of cultural issues was included, using examples of
the impact of culture on DV, health care, and the screening process. The workshop
concluded with a discussion of screening and assessment techniques, as well as
intervention and referral resources. An outline of the presentation is provided in
Appendix A. Attendees were also provided with handouts that allowed them to follow
along with the didactic components and elaborated on information that was introduced
during the workshop. The training workshop provided a more extensive coverage of DV
issues than the staff had previously received. Furthermore, they were introduced to the
new screening tool that would be used with patients and the presenters modeled the
screening and discussed different ways of asking the screening questions and responding
to disclosures.

A total of 22 health care providers, including residents, nursing staff, and social
workers, completed surveys prior to and/or following the DV training. Seventeen
providers completed pretests and 16 completed posttests. Of these, 11 providers
completed both a pre-training and post-training survey that could be matched in data
analyses. There were six nurses, twelve doctors, two social workers, and two medical
assistants who completed surveys overall. Eight physicians, two social workers, and one
nurse completed the 11 surveys that could be matched. Overall, most respondents were
female (77.3%) and Caucasian (81.8%), while 9.1% were African American, 4.5% Native

American, and 4.5% Asian/Pacific Islander. The mean age was 37.15 (S_D = 10.19) and

58

the mean number of years spent working in an obstetrical/gynecological clinic was 5.34
($2 = 6.61). There were three residents in each residency year (first, second, third, and
fourth) who responded, which represents the full residency program.

A resource manual was also developed for distribution to clinic staff providing
information on domestic violence and related issues, including community resources.
There were eight manuals given to the clinic the day of the provider training, to ensure
that a manual would be available to each staff person at any given time. This manual also
contained information expanding on that presented in the training. Because the training
was relatively brief, it was not possible to cover all relevant information in-depth.
Therefore, the manual was intended to serve as a supplement to the training, to possibly
address specific questions that might arise with individual patients or simply to inform an
interested provider.

Following the training workshop, the researcher met with the residents three times
to discuss problems they were having with the screening procedure and to address
questions and concerns. The residents also provided feedback on their experiences with
the screening tool, to the researcher and to each other. Issues such as screening technique
and necessary frequency were discussed. Thus, these meetings served as “boOster
sessions” to the training workshop and provided a time for residents to discuss specific
issues about DV screening that they encountered after beginning to screen patients.

Patient Surveys. After the universal screening procedure had been in place for

 

approximately six months, patients completed anonymous surveys that provided evidence
of the potential barriers to DV disclosure in the health care setting, as well as providing

feedback on the screening process as it was instituted in the clinic. Patients were

59

approached by a research assistant while in the waiting area at the clinic and asked if they
would like to participate in the study. A cover sheet explained that completion was
completely voluntary and returning a completed survey implied consent. Completed
surveys were either returned to the research assistant or given to a clinic staff member to
return to the researcher. Participants were also given the researcher’s phone number in the
event that they had any questions concerning the study.

Ninety-nine patients completed this survey regarding DV screening in the health
care setting. Both pregnant and non-pregnant patients over age 18 were included in the
recruitment process. Eleven additional women who were approached refused to
participate in this portion of the study. Of those completing a survey, the mean age was
27.67 years @ = 9.09). Most of the respondents were single (58.2%), 31.6% were
married, 8.2% divorced, and 2.0% were separated. The racial/ethnic distribution is
believed to accurately reflect the clinic population: 56.6% were Caucasian, 23.2%
African American, 9.1% Hispanic, 6.0% biracial/multiracial, 3.0% Asian/Pacific Islander,
and 1% (1 patient) each were Native American and Middle Eastern.

Income was not assessed because this clinic is known to serve a low-income
population. In terms of educational attainment, 13.1% had less than a high school
education, 38.4% were high school graduates, 9.1% had obtained a GED, 25.3% had
attended some college, and 12.1% were college graduates. This information was not
reported by two respondents. Just over half (51.5%) of the respondents were pregnant at
the time they completed the survey. Of these women, 16% were in the first trimester,

30% in the second, and 54% in the third trimester of pregnancy. The mean number of

60

prenatal appointments the pregnant women reported having had at the time of the survey
was 9.4 (S_D = 8.23).

Intervention. Women screening positive for DV were referred for additional
intervention with a clinic social worker. Doctors were asked to indicate on the AAS if a
referral was made, although social worker referral for non-medical issues was a regular
part of the clinic procedures prior to implementation of this project. Thus, the
intervention component was not new to the clinic and did not require additional training
of clinic staff to complete.

Measures

Patient Survey. The patient survey was a brief, anonymous, self-report measure

 

administered six to eight months after provider training. It assessed basic demographic
information and comfort in discussing personal information with health care providers.
Patients reported their comfort in discussing DV with specific healthcare professionals,
barriers that might affect their comfort such as provider gender and race, and if they
would disclose abuse to their physicians. Patients were also asked questions about
potential barriers to a trusting patient-provider relationship, ranging from attitudes about
DV and knowledge of resources to feelings about the doctor-patient relationship. Most
questions were close-ended, with forced choice attitude questions having a 5-point scale
from “strongly disagree” to “strongly agree.” One question asked patients to list up to
five potential barriers to DV disclosure in an open-ended format. These responses were
coded according to the three-level model (intrapersonal, interpersonal, and structural) of
potential barriers to DV screening and disclosure. Some survey questions were taken

from Hayden, Barton, and Hayden (1997). See Appendix B for a copy of this survey.

61

Provider Survey. The provider survey was given to all clinic staff prior to and six

 

months after the DV training session. This questionnaire assessed knowledge, attitudes,
perceived barriers, and behaviors related to domestic violence in general and specific to
pregnancy, as well as questions specific to cultural issues. Knowledge and attitude
questions were presented with a 5-point scale from “strongly disagree” to “strongly
agree,” and questions pertaining to potential barriers were also on a 5-point scale, fiom
“not important” to “very important.” Also included were questions indicating the
frequency with which certain screening behaviors occur. This survey was adapted from a
number of existing surveys, including one developed by the Department of Obstetrics and
Gynecology of the College of Human Medicine at Michigan State University (P.
Vasilenko, personal communication, March 1, 1999) and a survey developed by ACOG
(see Horan et al., 1998). The pre- and post-training surveys were matched, using a four-
digit code number as an identifier.

After analyzing the pre-training surveys, minor revisions were made that appeared
on the follow-up surveys. First, certain questions were eliminated because it was
determined that there was no connection between the item and the training, so change
would not be expected based on the educational intervention. Second, several questions in
the “attitudes/opinions” section were eliminated because of a ceiling effect prohibiting
any significant change from taking place. Finally, the section of the survey assessing
barriers to DV screening was enhanced by the addition of two closed-ended items and
two open-ended items soliciting additional barriers as well as suggestions for more

effective DV screening. This combination of deletions and additions to the survey kept it

62

a reasonable length to maximize provider completion. The pre-training survey is in
Appendix C and the post-training survey is in Appendix D.

Some items on the survey were condensed into factors, while others were kept as
individual items, due to the importance of their individual content. Some items were
reverse-scored so higher agreement scores were desired for all items. The
knowledge/attitude items were factor analyzed, resulting in six factors. Because the
factors that were produced did not appear to have any substantive meaning, factors were
recreated based on inter-item and item-total correlations and item content, and reliability
analyses were conducted to determine if they were statistically acceptable. Based on this
process, the knowledge/attitude questions were condensed into three factors: general
knowledge about DV (3 items), personal competence/responsibility (4 items), and
barriers to screening (6 items). One item (Patients have a right to privacy about family
matters like abuse) was eliminated due to its lack of relation to any subscale. Reliability
analyses were conducted separately for the pre- and post-training surveys, the results of
which appear in Table 2. The relatively low reliability coefficient found for General DV
Knowledge on the post-training survey was primarily due to the fact that this factor only
consisted of three items, making it sensitive to small fluctuations in item means that led
to lower inter-item correlations. The mean of each item did increased fi'om pre- to post-
test; however, inter-item correlations and item-total correlations decreased on the post-
training survey.

The section of the provider survey directly assessing the importance of specific

variables as barriers to DV screening was kept separate from the knowledge items that

63

indirectly assessed potential screening barriers. These items were analyzed both as a sum

score and as individual factors impacting screening.

Table 2

Cronbach’s Alpha Reliability Coefficients for Provider Survey Factors

 

 

Factor Pre-training Survey Post-training
Survey
General DV Knowledge .7417 .3138
Personal Competence/Responsibility .7605 .6549
Barriers to Screening .7350 .7312

 

 

 

 

 

Abuse Assessment Screen. This screening instrument consists of five questions

 

assessing emotional, verbal, physical, and sexual abuse, and is a modified version of the
Abuse Assessment Screen (AAS) developed by McFarlane, Parker, Soeken, & Bullock
(1992). Their instrument was found to provide valid and specific identification of abuse
in a sample of pregnant women. For each question that was answered “yes,” the patient
was also asked if the act occurred in the current pregnancy and the last time it happened.
There were also questions assessing when any abuse began and who committed the
abuse. There was space for screening at 10 visits, although no patients were screened
more than five times. At the bottom of the instrument was a place for screener initials and

an indication if a social worker referral was made. Because patient race or ethnicity is not

64

a regular part of the medical chart, it was added to the AAS for research purposes. A copy
of the AAS was placed in every woman’s chart, with the intent that the physician would
complete it at each prenatal visit. The instrument was in carbon copy format, so at the
time of the second chart review, the carbon copy was pulled from each eligible patient’s
chart and attached to the other chart data collection form, to facilitate data entry. There
was a variable included in the database to indicate if the AAS was used, in the chart but

not used, or not present in the chart. See Appendix E for a copy of the AAS.

65

RESULTS

The following results are organized according to the four primary components of
the project: archival data collection (medical chart review), provider training and surveys,
patient surveys, and social worker intervention.

Archival Data Collection

 

Following provider education and training, all pregnant patients were to be
screened for DV throughout their pregnancies. It was believed that this procedure would
lead to a significant increase in the number of times women were screened and the rate of
disclosure of violence, compared with the existing clinic screening procedures.
Exploratory analyses were conducted to determine if there were differences in overall
screening or disclosure rates across racial/ethnic groups, to ascertain if race/ethnicity
serves as a barrier to effective DV screening.

Initial Chart Review. Because of the multiple screening methods used by the

 

clinic, patients could be missed on one and still screened through another. Of the 300
charts reviewed, there were 10 patients (3.3%) who received no screening at all,
according to chart forms. In some cases these forms were missing, in others they were in
the chart but left blank, and in other cases it was noted that the DV questions were not
asked because the partner was present during the appointment.

Of those patients screened for abuse, 12.4% reported physical abuse on the written
screening form, 2.8% reported sexual abuse on the written screening form, 16.3%
verbally reported physical abuse at the nurse screening, and 28% reported some type of

abuse to the social worker. The written and nurse screening forms specifically assess for

66

DV during the past year, whereas the social workers frequently include childhood or
other past abuse on their reporting form. Thus, these data illustrate a greater rate of
reported abuse when the assessment is verbal (versus written) and screening questions are
not specific to type or recency of abuse. These results support previous research that
found a significant increase in reported prevalence of abuse when asked directly in a
verbal interview, versus self-report (McFarlane, Christoffel, Bateman, Miller, & Bullock,
1991)

Second Chart Review. A total of 272 charts were reviewed post-provider training.

 

Each chart was reviewed at least six weeks postpartum, to allow for the postpartum visit
screening to be included in the analyses. All patients who gave birth after the provider
training and had seen a physician at least once were included in the analyses. Ten charts
were eliminated from the analyses because the patient had not actually had an
appointment with a doctor prior to giving birth, negating the possibility of physician DV
screening from taking place.

In addition to the physicians using the AAS, the nurse and social worker
screenings continued. A total of six patients (2.2%) were not screened at all through these
methods. Of those patients screened for DV by the “traditional” methods, 10.7% reported
physical abuse and 0.8% reported sexual abuse on the written screening. The nurse verbal
screening had a disclosure rate of 9.9%, and the social worker note indicated 27.1% of
patients screened had experienced abuse. Because the clinic protocol for written and
verbal nurse and social worker screening did not change post-training, it was not expected
that these screening or disclosure rates would change significantly from the initial chart

review.

67

A total of 46.9% of those patients seen during the reviewed period were screened
by their physician at least once, as recorded on the AAS. There was a significant increase
in physician screening during the project, as compared to the pre-training physician
screening rate of zero (p < .0001). Of those not screened by a doctor, 25.7% of charts had
the AAS but it was not used, and 28.3% were missing the AAS. A total of 87 patients
(33.2% of charts reviewed) were screened once, 23 (8.8%) were screened twice, 12
(4.6%) were screened three times, and only one patient was screened five times. The
mean number of physician screenings was 0.66 (_S_D = .87). In all but one case, if the
screening occurred at all, all four of the abuse questions (fear, verbal, physical, and
sexual) were asked.

In some cases, the AAS form was used to report that the screening was not done
due to the presence of the partner. In two of these cases, this situation prevented any
screening from occurring, whereas in other cases at least one screening did occur in the
partner’s. absence. Of the six patients who were not screened by any of the traditional
clinic procedures, two were screened by their physicians. The addition of the physician
screening did not significantly increase the number of patients screened overall, because
of the small sample who were not screened by a nurse or social worker. However, there
would be clinical significance in identifying even one more abused woman that was
missed by the traditional screening procedures.

A total of 37 patients, or 30.1% of those screened, disclosed some type of abuse to
their doctors. At the first screening, 0.8% reported fearing someone in their home, 15.4%
reported verbal abuse, 19.5% physical abuse, and 7.3% sexual abuse. Because of the

decrease in number of women screened and, consequently, the number of disclosures,

68

beyond one visit, these reported percentages were used as estimated rates of different
types of abuse in this patient population. These results fall within the range of other
estimates of lifetime prevalence of physical abuse in similar populations (Gazmararian et
al., 1996). These other studies examining violence during pregnancy did not specifically
look at prevalence of non-physical forms of abuse.

Information regarding the recency of abuse and relationship to the perpetrator
differed somewhat between nurse/social worker notes (in the second chart review) and
information obtained from the AAS. According to the regular chart notes taken by a nurse
and/or social worker, of those patients reporting abuse, 25.3% were abused within the
past year, 9.6% were known to have been abused during the current pregnancy, 19.4%
were abused 1-3 years ago, and 28.9% reported abuse more than 3 years ago. This
information was not available for 16.9% of patients reporting abuse. Based on the first
AAS screening, the majority (47.0%) of patients disclosing abuse reported abuse
occurring more than 3 years ago. Only 8.8% reported abuse within the past year, 2.9%
were known to be abused during pregnancy, and 3.6% reported abuse 1-3 years ago. This
information was missing for 17.7% of reported cases of abuse.

The perpetrator, based on nurse or social worker notes, was primarily a current
(24.1%) or former (38.6%) partner. In 9.6% of cases the abuser was a parent, 7.2% were a
family member other than parent or sibling, a friend or acquaintance in 3.6% of cases, a
stranger in 3.6% of cases, and a sibling in 1.2% of cases reported. This information was
missing for 12.0% of those patients reporting abuse. According to AAS data for the first
screening, 27.8% of abuse was perpetrated by a current partner, and 47.2% by a former

partner. A parent was reported in 8.3% of cases, other family member in 5.6%, and a

69

stranger in 2.8% of those cases reported. Identification of perpetrator was missing in
8.3% of cases.

Analyses were conducted to determine if race/ethnicity impacted DV screening
and/or disclosure. Based on one-way ANOVA, abuse disclosure did not differ
significantly across race. However, there was a significant difference across racial groups
in the number of times a patient was screened, indicated by AN OVA. Post Hoc tests
revealed that Caucasian women were screened more times than women in the “other”
category, E (4, 193) = 2.836, p < .05. Table 3 illustrates this finding. When race was
condensed into a dichotomous variable, white women were screened significantly more
often than were women of color, 1 (196) = 2.614, p < .01. According to these findings,
race/ethnicity did serve as a barrier to DV screening but not to disclosure of abuse.
Therefore, it was more of a barrier to effective DV screening for physicians than for

patients.

70

Table 3

Number of DV Screenings by Patient Race/Ethnicity

 

Racial/Ethnic Group 9 Mean number of times screened SE

 

Caucasian 121 .942 .081
African American 40 .650 . 141
Hispanic/Latina 20 .600 .199
Asian American/ 8 1 .000 .3 14

Pacific Islander

Other 9 .000 .296

 

 

 

 

 

 

Note. 3 = sample size; SE = standard error.

It was hypothesized that women of color who disclosed abuse to their physicians
would do so at later visits than Caucasian women. It was believed that women of color
would require more time to develop trust in their providers, while both Caucasian women
and women of color would be equally likely to experience onset of abuse later in
pregnancy. However, only two women disclosed abuse after the first DV screening: one
patient disclosed the second time she was screened and another the third time she was
screened. The woman who disclosed at the second screening was abused the day before
the screening occurred, indicating late onset of abuse. The woman who disclosed the third
time she was asked by her physician had also disclosed one incident of abuse to the social
worker at her initial visit, and still only reported that one incident of abuse. No analyses

were conducted to determine the causes or dynamics of late disclosure, due to the small

71

incidence of such an event. In this patient sample, race did not appear to impact time of
disclosure.

The only abuse reported to a physician that had not been reported to either a nurse
or social worker during their screenings was past abuse (with the exception of the patient
discussed above). This finding is partly due to the wording of the written and verbal nurse
screening, which specify violence within the past year. The number of women who were
not screened for DV in any way did decrease after project implementation, from 3.3% to
1.5%, although it is not known if the project actually caused this result, as it was not
solely due to physician screening.

Providers’ race/ethnicity and gender were also examined to determine their impact
on overall patient disclosure rate. It was hypothesized that women would be more likely
to disclose abuse to providers of the same race/ethnicity as themselves. Exploratory
analyses were conducted to determine if provider gender impacted patient disclosure. In
order to examine the impact of providers’ race and gender on patient DV disclosure, two
dichotomous variables were created to indicate same/different race and same/different
gender. The analysis for same/different race was based on 198 patients, due to missing
race/ethnicity data. A Fisher’s Exact Test indicated that provider race was a significant
factor in patient disclosure, with more patients disclosing to providers of the same race as
themselves (p < .05, l-sided). There was no significant effect of provider gender on

patient disclosure, 760, N = 262) = 2.116, p > .05 .

72

Provider Training

 

It was expected that domestic violence education would reduce barriers to
conducting DV screening by increasing providers’ knowledge of DV and awareness of
cultural issues related to DV, changing attitudes towards DV and its victims, decreasing
the number and impact of perceived barriers to conducting universal DV screening, and
increasing the amount of self-reported DV screening as measured six months afier
training.

Figure 2 illustrates how the training workshop was predicted to impact the
providers, in terms of the theoretical barriers model to DV screening. The training was
expected to primarily affect barriers at the intrapersonal level, by increasing knowledge
and changing attitudes. The training was also expected to improve providers’
interpersonal skills in discussing abuse with their patients, as well as raising their
awareness of the impact of their screening behavior. Although the workshop was not
designed to actually change any barriers at the structural level, it had the potential to
indirectly affect these barriers by changing perceptions and raising awareness of their

impact.

73

 

 

Structural ,_ _ __ _. - ,7 7 -- Training

 

 

Interpersonal

Intrapersonal

   

Provider

j > Screening

 

Figure 2. Direct and indirect effects of training on provider barriers to domestic violence

screening.

There were 17 pre-training surveys completed by 9 physicians, 2 social workers, 4
nurses, and 2 medical assistants. The follow-up surveys were completed by 11
physicians, 2 social workers, and 3 nurses. There were matched surveys available for 11
respondents: 8 physicians, 2 social workers, and 1 nurse.

Based on the matched sample of 11 providers, significant positive differences
were found for the following items/factors: 1) How confident do you feel about your skills
in asking patients about domestic violence; 2) How fiequently do you ask direct, specific
questions about abuse; and 3) Competence/responsibility in addressing DV with patients
(p < .05 for all differences). When only the matched surveys for the eight doctors were

examined, confidence in screening skills and the competence/responsibility factor were

74

&

significantly higher at follow-up (9 < .05). The barriers to screening factor was significant
lower (p < .05), indicating physicians’ perceptions of potential barriers had changed in a
positive direction. Therefore, the training workshop and additional experience with DV
screening increased physician’s sense of competence and responsibility in screening for
abuse, and decreased the perceived barriers present to screening. Their self-reported rate
of screening did not increase, in contrast to the actual AAS data.

Items from the “barriers” section of the survey were summed to create a total
score of the importance of barriers in DV screening. Repeated measures ANOVA was
planned, but due to the small sample size, between subjects analyses could not be reliably
conducted. Using paired samples t—tests, some significant differences were found between
the pre-training survey and the follow-up survey.

First, there was no significant difference in the providers’ ratings of the total
importance of barriers in their decision to screen for UV before and after training.
However, it is interesting to examine the specific barriers that were identified as having
some importance. The barriers were rated according to the following scale: 1= not
important; 2 = slightly important; 3 = somewhat important; 4 = quite important; and 5 =
very important. On the pre-training survey, time constraints was given the highest rating,
but the mean was only 2.59 (S_D = 1.18). Of all provider responses on the follow-up
survey Q1 =15), several barriers received mean ratings of greater than 3.00: language
barriers M = 3.60, S_[)_ = 1.12), time constraints (M = 3.33, S_I_)_ = 1.05), and the presence
of the patient's partner (M = 3.33, S_D_ = 0.98). The 11 physicians responding rated the
presence of the patient’s partner as the most important barrier to screening M = 3.55, §I_)

= .93). Thus, the overall ratings of the importance of these barriers to screening actually

75

increased at follow-up; although it was not a significant change, it was in theopposite
direction as expected. This finding is also in contrast with the decreased endorsement of
barriers in the “knowledge” items on the survey.

On the open-ended item asking providers to list up to five reasons why they may
have had difficulty screening all patients for DV, few responses were given. Those
responses that were provided primarily reiterated the close-ended section, such as time,
partners unwilling to leave the room, and language barriers with the partner as translator.
One provider noted that some patients are only seen once before delivery and if the
partner is present at that time, screening will not occur. In a high-risk patient population
where late (or no) prenatal care is common, this issue is likely to be more prevalent.
Another response noted that sometimes patients will reveal more once you have
developed a relationship with them, which. recognizes the importance of the interpersonal
relationship in effective DV screening.

There was minimal change in providers’ knowledge of institutional policies
following training. Prior to training, 59% of respondents were aware that there is a
written protocol for patient DV screening; 6% believed there was not, and 35% did not
know. On the follow-up survey, 62.5% of respondents knew of the written protocol,
while 18.8% did not know. However, the number who replied that there is no written
protocol increased to 18.8% of respondents. No significant relation was found between
the pretest and posttest responses on this item. Since that written policy was discussed in
the training session, this finding was surprising. The number of providers who stated they
conduct a one-on-one discussion with each patient increased from 76.5% to 87.5%,

although this does not represent a significant change.

76

In providing suggestions for incorporating DV screening into prenatal care,
responses indicated that once it becomes routine, it becomes easier—both in terms of the
providers’ ability and in explaining to patients why it is being done. Longer-term
evaluation of this project would be necessary to determine if this response was reflected
in actual screening behavior over time. Finally, in evaluating the training session, 80% of
respondents agreed/strongly agreed that it was helpful. In addition, 93.7% of those
responding indicated that they are increasingly comfortable asking patients about partner
violence; only one provider disagreed with this statement. These responses support the
value of the DV training workshop.

Survey of Patients

 

Clinic patients were surveyed to provide their perspective on instituting an
effective universal DV screening protocol. It was expected that overall, patients would
support the concept of universal screening and report feeling comfortable with the
discussion, but that women from minority groups would report more barriers to
disclosure and less overall comfort in disclosing abuse to a health care provider.

First, patients’ views on what behaviors define domestic violence were assessed,
producing information that is important to ”consider in implementing screening
procedures. While most women reported acts of physical abuse as examples of domestic
violence, a number did not identify emotional, verbal, or sexual abuse as acts of domestic
violence. These results emphasize the need to include specific assessment questions when
screening for abuse. Table 4 provides the percentage of respondents that identified each

statement as describing domestic violence. .

77

Table 4

Patients’ Definitions of Domestic Violence

 

 

 

STATEMENT % ENDORSED

Pushing, shoving, slapping, punching, kicking, choking 96.0
Assault with a weapon 89.9
Holding, tying down, or restraining you 89.9
Threats of harm, or intimidating you 83.8
Keeping you from seeing your friends or family 65.7
Degrading, humiliating, or making you feel bad about yourself 72.7
Trying to make you perform sex acts against your will 85.9
Hurting you physically during sex or assaulting your genitalia 84.8
Forcing you to have sex without protection against pregnancy or 80.8
sexually transmitted diseases

 

 

 

 

Patients were asked, on a scale of 1-5 (very uncomfortable to very comfortable)
how comfortable they would be discussing abuse with a nurse, doctor, clinic social
worker, and community counselor. The means were slightly above “not sure,” with no
significant differences between provider types: 3.14 @ = 1.55), 3.24 LSD = 1.61), 3.19
(SD = 1.56), and 3.22 QT; = 1.52), respectively. These averages are partly a result of U-
shaped distributions on each response. In reporting how comfortable they would be
talking with a nurse, 25.0% responded very uncomfortable, 12.5% somewhat
uncomfortable, 11.5% not sure, 26.0% somewhat comfortable, and 25.0% very

comfortable. For comfort with doctor, 25.8% reported very uncomfortable, 9.3%

78

somewhat uncomfortable, 12.4% not sure, 20.6% somewhat comfortable, and 32.0% very
comfortable. For comfort with a clinic social worker, 24.0% responded very
uncomfortable, 12.5% somewhat uncomfortable, 12.5% not sure, 22.9% somewhat
comfortable, and 28.1% very comfortable. Finally, for comfort in talking with a counselor
in the community, 22.1% reported they would be very uncomfortable, 12.6% somewhat
uncomfortable, 13.7% not sure, 24.2% somewhat comfortable, and 27.4% very
comfortable. Figure 3 illustrates the distribution of responses regarding comfort in

discussing abuse with a physician.

 

40

ereent of Respondents

 

 

 

' Very Uncomfortable Not ure Very Comfortable

Comfort with Doctor

Figure 3. Patients’ Comfort in Discussing DV with their Physicians.

Over half (57.6%) of respondents reported that their doctor had asked them about

DV, and this proportion increased to 76.5% when only considering pregnant patients,

79

whom the screening was targeting. These percentages are substantially higher than those
screened according to AAS data (46.9% were screened with the AAS). It is possible that
some physicians did discuss DV with their patients without completing the AAS form,
which was seen by some as inconvenient additional paperwork. Alternatively, patients’
recollections of DV screening may simply have differed from the actual occurrence of
formal screening.

There were two questions assessing whether or not the respondent would disclose
abuse to her doctor, if asked. One question asked for a “yes” or “no” response, while the
other was answered on a 5-point agreement scale. Of those responding to a yes/no
question, 89.2% reported they would tell their doctor if they were being abused, if the
doctor asked. Six respondents did not answer this question. On the agreement scale,
70.1% indicated they would tell their doctor if they were being abused (strongly agree or
agree), while only 7.2% stated they would not do so (strongly disagree or disagree).
Another 22.7% responded they were unsure. These data support previous research
indicating that women will disclose abuse if asked directly.

On a 5-point scale of agreement, most patients (71.5%) reported being
comfortable discussing personal issues with their doctor. Approximately 67% of
respondents stated they are more comfortable discussing abuse with a female doctor,
while only 20.4% are more comfortable discussing abuse with a male doctor. These
results do not indicate that patients will not discuss abuse with a male doctor, but that
fewer prefer a male doctor over a female doctor. In reporting if they are more comfortable
discussing DV with a doctor of the same race, 67.3% disagreed or strongly disagreed,

while 20.4% agreed or strongly agreed.

80

Several closed-ended questions assessed potential barriers to DV disclosure,
indicating that some items are not significant barriers as was believed. In response to the
statement, “I believe that domestic violence should not be discussed outside the family,”
83% strongly disagreed or disagreed, and only four patients responded in agreement. The
majority (88.6%) of patients also disagreed with the idea that they would be offended if
their doctor asked them about domestic violence. Another statement read, “I would be
concerned about discussing abuse with my doctor because of how my partner would
react.” Of those responding, 66.7% strongly disagreed or disagreed, while 18.7% agreed
or strongly agreed. It is possible that more women did not identify this item as a barrier,
because such a concern might not occur to a non-abused woman. It is likely that the
percent who agreed would be greater in a sample of abused women. Concern about being
treated unfairly by the police as a barrier to disclosure was also assessed, with only 21%
of respondents agreeing that this was a concern. Surprisingly, 82.4% of respondents
indicated they would know where to go for help if they were being abused, implying that
appropriate referrals are not the greatest need for these women.

Analyses were conducted to determine if patient demographics impacted
responses on the survey. For the analyses involving race/ethnicity, the revised categories
of race were used in order to increase cell size; the biracial and multiracial categories
were combined, and the Native American and Middle Eastern categories were grouped
with “other.” Because some of these cells remained rather small, some analyses were also
conducted with a dichotomous race variable (white/non-white).

Analysis of variance indicated that Asian/Pacific Islander women were

significantly more likely to respond that they felt more comfortable with a doctor who

81

was the same race/ethnicity as themselves, as compared to white women, F (5, 92) =
2.372, p < .05. However, this result was based on a group of three Asian women. When a
t-test was run with the dichotomous race variable, women of color were not significantly
more likely to endorse this statement, although the difference neared significance, 3 (96) =
-1.976, p < .10. There were no significant differences between racial groups on the 5-
point scale item regarding disclosing abuse to a doctor; however, there was a significant
difference in the dichotomous yes/no question about disclosing abuse between white and
non-white women, with white women reporting they would be more likely to disclose
abuse, 980, N = 93) = 6.667, p < .01. This finding is inconsistent with the AAS data, in
which race/ethnicity did not affect rate of abuse disclosure. The impact of race should be
considered by health care providers whose screening behavior may also be impacted by
patient race; findings suggest that Caucasian women are more likely to be identified as
abused, possibly leaving women of color without services.

The only relation that patient age appeared to have with the domestic violence
questions was the item regarding fair police treatment if DV was reported. Age was
positively correlated with this item (9 < .05), indicating that older women feared unfair
treatment more than younger women did. It was also found that marital status was related
to this item, with separated women also more likely to anticipate unfair treatment, I: (3,
90) = 2.879, p < .05. However, only two women were in the “separated” group, and they
were above the mean age, indicating a possible spurious relationship. There was also a

relation found between education level and comfort discussing personal issues with one’s

82

doctor, but post hoc tests revealed no significant difference between groups, due to lack
of power.

Barriers to Disclosing Domestic Violence. On the survey, patients were asked to

 

list up to five reasons why they might not tell their doctor if they were being abused.
Forty respondents listed at least one barrier in this section. A total of 122 responses were
provided that could be coded into one of the three levels in the theoretical barriers model:
intrapersonal, interpersonal, and structural. Several respondents wrote that they would
definitely tell their doctor or they had no reason to not tell, and these responses were not
coded as barriers. Table 5 provides a smnmary of responses within each level of the
model. Items at the intrapersonal level included feelings and other internal states.
Interpersonal responses included items specific to the doctor-patient interaction or
consequences from the doctor, as well as references to other relationships, such as
consequences from the batterer. Items were coded in the structural category if they either
referred to a system, such as child protective services or the police, or something specific
to the clinic setting, such as the presence of others during screening. Most of the
responses (76) fell in the interpersonal level, with 32 responses at the intrapersonal level.
Structural barriers were the least frequently reported, with 14 responses in that category.
The most frequent response overall was concern about the consequences of telling the
doctor. In many cases, the fear was from not knowing what the consequences would be,
which is a barrier easily reduced by improved communication. Fear of repercussions fi'om
the abuser was also a common response in the interpersonal category. The most prevalent

response in the intrapersonal category was shame/embarrassment.

83

For the 40 women who responded to the open-ended barriers question, analyses
were conducted to determine if race/ethnicity impacted the number or type of barriers that
were identified. Analyses were first conducted across race/ethnicity categories using
AN OVA, finding no significant differences. Because of the small sample sizes in some
ethnic groups, additional analyses were conducted withfthe dichotomous race variable to
examine differences between white and non-white respondents. There were no significant
differences between racial groups in the number of barriers listed within each category

(intrapersonal, interpersonal, and structural), or in the total number of barriers listed.

84

Table 5

Barriers to Disclosing Domestic Violence to a Physician

 

 

 

Depending on the kind of abuse
I can handle things on my own
You love the abuser

You don't care

You're pregnant by the abuser
Wouldn‘t want to feel guilty
Wanting to deny it was

happening

 

 

 

listen (4)

They might look at me, treat me/spouse
differently (4)

Negative response (2)

It wouldn’t help (2)

Interactions with Doctor (6)

Doctors are not concerned, are always in
a hurry (2)

Don’t know if you can trust him/her (2)
Doesn't make me feel comfortable
talking to him

If doctor was a jerk

 

Consequences from Partner (19)
Fear of partner finding out (7)
Situation might get worse (3)
Threats from partner (3)

Partner might try to kill me (2)

Fear of mate hurting family for spite
Being kicked out of home

In danger if I told

If I was dead

Other

Fear (15)

Don’t want anyone to know (9)
What would people (family/friends)
think? (5)

Afraid for your children (2)

 

 

INTRAPERSONAL INTERPERSONAL STRUCTURAL
Shame/Embarrassment(18) Consequences from Doctor/Staff (20) Fear of children being
Discomfort (3) Who doctor might tell, what they might taken away (6)

Don’t want to be alone (3) do (8)
Sadness They might not believe, understand, or If others (partner,

children, partner’s
family) were present (5)

Don’t want police
involvement/concem
about getting in trouble

(3)

 

Note: Numbers in parentheses are the number of times a response was given; one

response was given if there is no number listed.

85

 

Intervention

 

When women disclosed abuse to their physicians, they were to be referred to a
social worker for further intervention and community referrals. The patient was either
seen immediately after her prenatal appointment, or an appointment was made for a later
date at the time she checked out. As mentioned, a total of 37 women disclosed DV when
screened by a physician. At the first screening, 15 referrals were made. Of these, 11 were
to the clinic social workers, 2 were to other community agencies, and 2 were unknown.
At the second screening, three referrals were made, all to a clinic social worker. Only one
referral was made at the third screening, which was also to a social worker.

In some cases in which no referral was made according to the AAS, the doctor
noted on the AAS that a referral was not made because the patient refused it. This was
often the case when the abuse was in the past, suggesting that the patient felt she no
longer needed counseling or other intervention. Because the AAS did not specifically
assess reasons for not making a referral, it is not possible to account for all of the patients
who disclosed DV who did not receive a referral. It is also not possible with the available
data to determine how often physicians offered referrals that were refused, versus how
often the physicians failed to make an appropriate referral. Verbal communication with
the clinic social workers revealed that the only situations where a patient was not seen
immediately after disclosing abuse did involve past abuse, and where there was no
immediate need for intervention. In those cases, the social worker made contact with the

patient later, to ensure that no intervention was needed.

86

Because the social workers were already on-site and doctors regularly made
referrals to them for psychosocial issues, this component of the project did not require
behavior change by clinic staff. In a setting where this intervention system was not
already in place, additional training and evaluation would be necessary to ensure that
patients were receiving appropriate services. It would also be important to conduct
additional follow-up with those women who refused or did not receive referrals, to ensure

that no intervention was needed.

87

DISCUSSION

Health care providers in the Obstetrics and Gynecology Residency Program Clinic
at Sparrow Hospital received DV education and were trained to do universal screening for
domestic violence. This project expanded the current clinic training and procedures by
providing more in-depth DV education and encouraging the physicians to screen every
woman at every prenatal visit, versus the nurses and social workers only screening a
patient at her initial prenatal visit. The focus was on decreasing provider barriers to
screening. The project was not intended to impact patients directly. Rather, it was
believed that the impact on reducing provider barriers and improving screening behavior
would indirectly impact patient barriers and increase disclosure. The project was
evaluated through four main components: 1) archival data collection through medical
chart review; 2) provider training evaluated through pre- and post-training surveys; 3)
patient surveys assessing their views on DV and comfort with discussing DV with health
care providers; and 4) providing social worker intervention for those women disclosing
abuse to their physicians.

Archival Data Collection

 

Prior to project implementation, all women seen in the clinic for obstetrical care
were screened for DV as part of their intake interview with a nurse, as well as with a
social worker if the patient qualified for the Maternal Support Services Program. Thus, it
was believed that screening at the initial visit was 100% and 0% at subsequent prenatal
visits. There is evidence that continued screening throughout pregnancy is necessary to

maximize identification of abused women (McFarlane et al., 1992). Whether because of

88

the need to build trust or because abuse did not begin until later in the pregnancy, it was
believed that DV screening at all prenatal visits would lead to a higher rate of disclosure.
The number of screenings prior to disclosure was used as an indication of trust-building
over time, while a question on the AAS would determine if there was late onset of DV
during the pregnancy. It was expected that lack of trust would present as a greater barrier
to DV disclosure for women of color than for white women.

The chart data indicate that a substantial number of patients had experienced, or
were currently experiencing, abuse by an intimate partner. Physician screening lead to the
greatest report of abuse (30.1%), partially because this screening included verbal,
physical, and sexual abuse, while the traditional screening methods focused on physical
and sexual abuse only. The large number of abused women that were being identified
prior to this project shows that this clinic did have commendable screening procedures in
place, although the physicians were not difectly involved. The overall number of missing
written or verbal nurse and social worker screening forms did not significantly change
over time, although the number of patients who were not screened at all did decrease. It is
possible that the researcher’s presence in the clinic served as a reminder to the other staff
to conduct the DV screening that had already been in place as clinic procedure. Had the
clinic not already had a DV screening protocol in place, the physician screening would
likely have identified a greater number of previously unidentified abused women than it
did, since close to half of all pregnant women were screened by a physician in the six-
month period of study.

Statistical comparisons across screening methods (physician versus “traditional”)

could not be made due to the different assessment forms used. However, there were some

89

noteworthy differences in the information obtained from each of the two methods. For
instance, there were differences concerning recency of abuse between the different
screening methods, although both rates do fall within the range found in other studies of
violence during pregnancy (Gazmararian et al., 1996). It is not known why there were
such differences in reporting of the last occurrence of abuse. One possible reason why the
prevalence of abuse in the past year was greater for the “traditional” screening is the way
in which this information was obtained from the records. The written forms (which are
reviewed verbally by a nurse) specifically ask if abuse has occurred within the past year.
If a patient answered “yes” to this query and no further information was available, the
patient was coded as having experienced abuse within the past year in the database. If
there was information in the chart concerning when the abuse occurred, it was sometimes
possible to determine if it was or was not during the current pregnancy. Therefore, this
information was not always obtained directly by asking the patient, as it was on the AAS
that was completed by the physicians. It is possible that fewer women verbally disclosed
recent abuse, although a nurse or social worker may have inferred it (either correctly or
incorrectly). It is also possible that women who had been abused longer than one year ago
responded “yes” to the written screening items, and were coded incorrectly as a result.
Because the physician screening did not specify abuse in the past year, physicians
identified more past abuse than the written screening forms, changing the proportions. It
is also expected that differences arise when several forms of abuse (emotional, verbal,
physical, and sexual) are grouped together, versus determining recency of abuse

specifically for each type of abuse.

90

Past abuse was more commonly reported than ongoing abuse, but this finding
does not lead to the conclusion that screening is not important. First of all, approximately
one-quarter of those women reporting abuse were abused by their current partner, leaving
the possibility of recidivism. Second, the effects of past abuse, particularly sexual abuse,
can continue to impact a woman’s health, including pregnancy outcome. Finally, there is
the possibility that women in currently abusive relationships did not disclose the abuse as
often as those women who were no longer being abused, due to barriers such as shame
and fear that might be more salient for a woman in an ongoing abusive relationship.

The impact of race/ethnicity on the patient-doctor interaction was also examined
through the chart data. It was hypothesized that having a provider that was a different
race and/or gender would be a barrier for some women during DV screening. Also of
interest was the potential for physicians’ biases in conducting DV screening, being either
more or less likely to screen for abuse in a minority patient.

Although the data presented here suggest that abuse disclosure is not dependent
on patient race, this conclusion is based on the assumption that there are equal rates of
intimate partner violence across racial and ethnic groups. Although this assumption has
been supported empirically (e.g. Cokkinides & Coker, 1998), the numerous problems in
estimating prevalence and incidence of violence prevent any final conclusions from being
drawn based on the available data. If it were known that in this patient population there
were equal rates of DV across racial/ethnic groups, then the AAS data would show that
patient race did not impact whether or not a patient disclosed abuse to her physician. One
should also consider the patient survey finding that white women reported being more

likely to disclose abuse, and interpret these data with caution.

91

The finding that white women were screened more fi'equently than women of
color was unexpected, and is an important finding to consider when working on the issue
of domestic violence and health care. It is possible that racism contributed to this finding;
for example, doctors may be more accepting of violence in minority groups and therefore
less likely to query about abuse. Alternatively, the doctors may have feared that their non-
white patients would perceive them as racist, i.e. only asking about abuse because of their
race. Another possibility is that language barriers contributed to the significant difference
in screening rates. Providers did not screen for DV if the patient did not speak English,
especially if her partner or another family member was serving as an interpreter. Even in
cases where the patient spoke some English, the language barrier may have dissuaded the
provider from conducting the screening. The possible explanations for this finding
include race as a barrier at both the intrapersonal and interpersonal levels, and even
potentially the structural level is there is institutionalized racism. Because of the
importance of this issue, further investigation should attempt to discern the cause of the
discrepancy in screening rates. Whatever the cause, health care providers need to be
aware of the impact of race on the DV screening process and the apparent tendency to
screen white women at greater rates than women of color. If race is a barrier to
identification of abuse, then women of color are less likely to receive appropriate
intervention.

It was not possible to examine the factors influencing disclosure after multiple
screenings, due to the small number of women that were screened multiple times and the
even smaller number that disclosed abuse after the first physician screening. There was

not a significant difference in the total number of visits between women who did disclose

92

abuse and those who did not, but the first AAS screening did not always occur at the first
physician visit. Some doctors did screen for DV at the first prenatal visit, while others did
not first screen a patient until later in her pregnancy. AAS use began with all pregnant
patients, regardless of the age of gestation, at the time of project implementation,
contributing to this inconsistency. Although physicians were asked to record the date of
each screening, this was rarely done. Therefore, there is not enough information available
in the chart records to firlly examine the issue of trust-building between doctor and
patient, and how that might impact DV screening and disclosure.

The chart data show that provider gender did not impact patient disclosure of DV.
However, provider race did affect disclosure, with more women disclosing abuse to
providers of the same race as themselves. This result may have been influenced by the
fact that 61% of the patient screened were white, as were 82% of the physicians that did
the screening. Thus, it was simply more likely that a patient and doctor would both be
white, and therefore more likely that a patient would disclose abuse to a physician of the
same race. However, this finding may also be a consequence of a racial barrier at the
interpersonal level.

While recommendations of direct questioning are the norm in the literature on DV
screening, some researchers have suggested that more indirect questioning may be useful,
such as when cultural differences are present, in order to seem less judgmental (Gerbert,
Caspers, Bronstone, Moe, & Abercrombie, 1999). Gerbert, Caspers, et al. also found that
screening rarely led to direct disclosure; rather, physicians reported patients hinting at
abuse, or showing signs in their body language or hesitation in responding. Gerbert,

Abercrombie, Caspers, Love, & Bronstone (1999) described a complex “dance” between

93

providers and patients, involving both direct and indirect communication by each party.
This dance sometimes led to identification of abuse, but the patients’ sense of being
validated by the provider was independent of identification, and was a vital step in the
process of seeking help. The experiences of physicians in the Gerbert, Abercrombie, et a1.
study led them to redefine the desired outcome of screening to be compassionate asking,
rather than disclosure. Showing concern and validating an abused woman’s experience in
a nonj udgmental manner may be a first step towards providing further intervention. Many
battered women will not disclose abuse, but knowing that their physician cares and is
available if they do choose to disclose at a later time may be considered a successfirl
intervention.

Because the current project focused on direct screening and disclosure, the subtle
signs or hints from patients may have been missed by the physicians, who were not
necessarily trained to recognize them. Likewise, the actual disclosure rate may not
represent all of the abused women helped by the physician screening. A true assessment
of the indirect communication that may have occurred between patients and physicians
during the screening process would be extremely difficult, if not impossible, to quantify.
Previous research on these subtle behaviors has relied on qualitative data and small
sample sizes. In this study, these data were not recorded in a systematic manner.

Overall, the results indicated that physician screening in addition to nurse and
social worker screening did not significantly increase the rate of DV disclosure. However,
training all clinic staff so that indirect questioning and validation can occur is important,
and the clinic should identify at least one group of professionals that will continue to

screen for DV throughout pregnancy. Future training of health care providers might also

94

include a discussion of the indirect forms of disclosure, as well as the importance of
screening in absence of a direct disclosure.

Provider Training

 

A 1.5-hour educational seminar was presented that covered general issues of DV
and its potential effects on pregnancy, cultural issues related to DV, and reducing barriers
to universal screening. The overriding theme of the training was reducing provider
barriers at the intrapersonal, interpersonal, and structural levels. It was expected that this
training would not only impact knowledge and attitudes, but also decrease the number of
perceived barriers to conducting universal DV screening and increase behaviors related to
DV assessment.

The true impact of the provider training, as assessed by the surveys, cannot be
completely discerned from the small number of matched surveys that were available for
analysis. However, the finding that providers’ comfort and confidence in screening for
abuse did increase over time is an indication of project success. Clinic staff indicated that
screening became easier as it became more routine and as they gained experience
discussing abuse with patients. They also reported that the training workshop was useful
to their practice. Thus, while their awareness of barriers to screening may have increased
over time, growing experience may have assisted these providers in identifying means of
reducing or circumventing these barriers. More extensive evaluation would need to be
conducted to determine if their screening effectiveness actually improved overtime as
well.

The providers’ increased ratings of the importance of several barriers were an

unexpected finding. It may be that the providers were more aware of the barriers to

95

screening once they increased their attempts to screen for DV. It is also interesting to note
that the most important factors affecting their decision to screen were those at the
structural level. The few responses by providers on the open-ended question regarding
reasons they may have not screened also point to the need for structural changes that
allow for a private discussion to occur between patient and doctor. It is also notable that
the knowledge items assessing perceived barriers showed a change in the opposite
direction. It may be that these items weie actually measuring different phenomena,
adding to the complexity in identifying barriers in the DV screening process.

The physicians’ primary barriers to screening throughout pregnancy—time and
presence of patients’ partners—remained barriers after the project’s completion. While
these physicians perceived time constraints to be an important barrier, some research
indicates that patients understand physicians’ busy schedules and do not expect a lot of
time from them. Rather, they desire direct questioning which acknowledges the abuse and
an appropriate referral for further intervention (Rodriguez et al., 1996). It does not require
a great deal of time to convey compassion and understanding for a patient. It may be that
a combination of structural/procedural changes and providing more intervention
resources for referral would reduce physicians’ perceived barriers to DV screening.

Several of the identified barriers are issues that need to be addressed at the
administrative level of the clinic. Systemic changes need to be made to actually reduce
the impact of these barriers. For example, clinic procedures need to be adjusted so
physicians have time to interact privately with their patients, without alienating or raising
the suspicions of the partners that accompany them to prenatal appointments. While

training and educational interventions may increase individual providers’ skills and

96

willingness to screen for DV, the structural barriers cannot be removed through a short-
tenn intervention such as the one presented here. Discussions did occur with the residents
concerning how they might deal with asking a patient’s partner to briefly leave the exam
room, but more extensive clinic changes are necessary to overcome the doctor’s hesitancy
to do so. If such a procedure were standard for all patients, it would save the physician
from having to find creative solutions on a case-by-case basis. Administrators can be
made aware that these structural barriers are among those inhibiting effective universal
DV screening and assisted in finding innovative ways to change the clinic structure to
reduce such barriers.

Patient Survey

 

Previous research indicates that women will frequently not initiate discussion
about DV with their physicians, but they often do want the physician to inquire about
abuse experiences and will disclose abuse if asked (Friedman et al., 1992). In the current
study, it was not known if the screening procedures that were being initiated by
physicians were welcomed by all patients. It was also not known what specific factors
might prevent a patient in this clinic from disclosing abuse to her physician. As a way to
inform this project, patients were surveyed six months after the universal screening
procedure had been initiated on their comfort with discussing abuse with healthcare
providers, and perceived barriers to open patient-physician discussion about DV. Surveys
were given after the new screening procedures had begun, rather than prior to their
implementation, so responses would reflect patients’ actual experiences with screening,

versus their beliefs about how they thought they would react to physician screening.

97

In addition to attitudes about DV screening, patients’ definitions of DV were
assessed. It was found that while most women identified acts of physical abuse as
domestic violence, acts of emotional, verbal, and sexual abuse were not identified as
abusive by as many women. These results support the need for multiple and specific
screening questions, such as those on the AAS. If patients are only asked if they have
been victims of abuse or violence, some patients may respond negatively when in fact
they have been subjected to non-physical acts of abuse. It is also important for health care
providers to ascertain whether the patient sees herself as abused or not, even if she does
disclose acts of violence.

As expected, the majority of patients indicated that they would disclose DV to
their physicians if they were asked. Although abuse history was not assessedin
respondents, previous research found that a history of physical abuse did not affect
likelihood to disclose or patient support for routine screening (Friedman et al., 1992). The
women in this clinic sample did not report a preference for a particular health care
provider to conduct the screening. It appears that if a woman is comfortable discussing
abuse at all, then she is comfortable with each provider to a similar degree. If a woman is
uncomfortable, then it does not matter who is conducting the screening. However, even a
woman who is uncomfortable may still disclose abuse, as evidenced by the majority of
patients indicating they would disclose abuse if asked, compared to the number reporting
discomfort. It may be that women are accustomed to discussing personal, sometimes
painful, issues with health care providers, despite any discomfort they may be

experiencing.

98

While only a small proportion of patients reported being more comfortable
discussing DV with doctors of the same race as themselves, there was some indication
that women of color endorsed this idea more than Caucasian women did, with a trend
towards significance. In addition, white women were more likely than women of color to
report that they would disclose abuse if asked by their doctors. This finding is in contrast
to the medical chart data, which found that race did not affect disclosure of abuse. As
mentioned, however, that conclusion is based on the assumption of equal preValence rates
of DV. The discrepant findings from the tWo data sources point to the need for more
research on this issue. Furthermore, health care professionals conducting DV screening
should be aware that race may act as a barrier to abuse disclosure for women of color.

It is also important to consider the setting when interpreting the results found
here. In a residency clinic, patients do not often choose their doctor. In a private setting,
where a woman has selected her physician, different barriers may be present. For
example, a woman of color who is more comfortable with a doctor of the same race may
choose such a doctor in private practice, eliminating race as a potential barrier to open
communication. Physicians in clinics such as the one studied here need to beaware that
barriers such as race may be present to a greater degree than in a private practice.

The potential barriers to DV disclosure reported by patients on the open-ended
portion of the survey add particularly important information to the scant literature in this
area. Few studies have examined barriers in the screening process from the patients’
perspective, focusing instead on health care professionals’ perspectives. In addition,
many studies have not included a theoretical rationale for explaining barriers to DV

screening and disclosure. Those studies that have addressed patient barriers frequently

99

focus on the interpersonal level, specifically examining doctor-patient communication
(e.g. Plichta et al., 1996). While the majority of barriers reported did fall at the
interpersonal level, the Barriers Model allowed for an examination of barriers at other
levels as well.

The results of the patient survey identified interpersonal barriers as the most
common barrier to disclosing DV. Within this level, fear of consequences from the doctor
or retaliation by the partner were the most frequently reported. Among the intrapersonal
barriers named, shame/embarrassment was the most common. Finally, structural barriers,
such as concern about societal systems intervening, were mentioned, although not as
frequently as barriers at the other two levels.

According to these responses, there are steps that providers can take to decrease
the impact of certain barriers. By explaining clearly to their patients what they will do in
the event of a disclosure, through open and sincere communication, physicians may help
to ease patients’ concern of the unknown and lead to increased disclosure. It seems that
being unsure of what will happen is preventing disclosure, and reassuring women of
confidentiality and explaining what services will be made available may lessen the effects
of this barrier. Patients also reported barriers that represent the realities of battered
women, such as justified fear of partner retaliation and shame in telling anyone about the
abuse. Understanding these realities may help providers to better communicate with their
patients. Although they cannot eliminate them as barriers, showing understanding,
validation, and respect may help a battered woman feel safer in disclosing the abuse.

In terms of culture, race/ethnicity was not specifically reported as a barrier to

disclosure by patients and no significant differences were found in responses across

100

patient race/ethnicity. However, other cultural issues can be seen in patient responses. For
example, the traditional culture of healthcare, including the power differential between
patients and doctors, is reflected in patients’ concern for consequences from the staff and
reports of negative interactions with their physicians. If a patient does not feel like she
has control over the information she shares with her physician, then she will likely
disclose less. Furthermore, fear of their children being taken away or other negative
societal involvement reflect the greater culture of our society, which continues to blame
women who are abused by their partners. Most of the respondents were low-income, and
these structural level responses may also reflect classism that they have experienced both
in and outside of the health care setting.

The patient survey responses can be used to improve DV screening practices. First
of all, providers can be made aware of these perspectives and trained in how to address
these barriers with their patients. Second, policies and procedures can be revised and
adapted to incorporate the needs of the patients. For example, women in this clinic
reported the presence of their partners as a barrier to disclosure (which was also a
common barrier reported by physicians). Clinic protocol can be modified to allow for
there to be a standard portion of the prenatal exam when doctor and patient are alone and
screening can take place. In addition, physicians in those states that have mandated
reporting of DV from health care providers should take into account that this policy may
decrease disclosure of abuse for fear of system involvement.

As mentioned, most research on DV screening in the health care setting examines
barriers from the provider’s perspective. These are certainly important barriers to address,

but it is also necessary to consider the patient’s perspective. In this study, the majority of

101

patient barriers were at the intrapersonal and interpersonal levels, while most reported
provider barriers were structural. The provider training addressed provider barriers at the
intrapersonal and interpersonal levels. Thus, only addressing barriers at one level alone
will not increase identification of abuse women. Removing the structural barriers may
assist doctors in doing the screening and may increase disclosure to some extent.
However, until medical professionals also address those barriers that the patient is facing,
there will continue to be women who are not comfortable disclosing their abuse
experiences. Considering the substantial impact of DV on women’s health, it seems that
all members of the medical community would want to maximize disclosure. This will
only happen if change addresses both providers’ and patients’ needs.

It is important to note the overlap between the three levels in the Barriers Model,
particularly between intrapersonal and interpersonal. In fact, arguments can be made to
shift some barriers from one category to the other. While this overlap makes quantifying
the barriers difficult, it is a good indication of the complexity of the screening and
disclosure process. In one recent study, providers identified patients’ unwillingness to
disclose the abuse, and other patient-related factors, as a primary barrier to effective
screening and intervention (Rodriguez, Bauer, McLoughlin, & Grumbach, 1999). Perhaps
being educated on the complex system of barriers to disclosure would help these
providers to understand patients’ hesitancy to disclose abuse when asked and prevent the
burnout and fi'ustration that some providers have expressed. While lack of disclosure or
patient change were not listed as barriers on the provider survey used in this study, these
were issues that arose in discussions with the physicians during training and follow-up.

Specifically, one physician expressed frustration at patients not taking referrals or leaving

102

their partners when intervention was attempted. The process of leaving was explained,
and providers were informed of examples where patients indicated that physician queries
sometimes led to change years later. In other words, the physicians were told not to rely
on immediate change as reinforcement for screening, but to hope that they were having an
impact that might be seen in the future.

Intervention

The clinic where this study took place is fortunate because they have two social
workers working on-site, who are available much of the time for patient intervention
when necessary. When a woman discloses abuse (or other psychosocial problems) to her
physician, she is referred to a social worker for further intervention and community
referrals. Although physicians were trained on providing some intervention and referrals,
it was believed that having the social workers available might increase the physicians’
comfort with raising issues about DV with their patients. At the very least, the social
workers’ presence eliminated concerns about how to deal with disclosure as a potential
barrier to screening.

The almost constant presence of a social worker in this clinic makes it rather
different than most health care settings. The finding that physicians did refer women
disclosing abuse to a social worker is positive, but not surprising. These physicians are
taught as part of their residency training that the social workers are available to assist
with patient care, and according to the social workers, they have no problem making
referrals and working in collaboration with them. Therefore, making referrals was likely
not due to the training component of this project, but is rather part of standard clinic

activities. The issue that needs to be addressed is what happens to abused women when

103

 

social workers, or a similar professional resource, are not available on-site and
immediately if needed. These residents may become dependent on the social workers’
presence and may not be prepared to deal with situations such as abuse when they are in
private practice. Furthermore, the question remains if physicians are even less likely to
screen for DV if there is no immediate referral resource available. This issue of
appropriate intervention should be a component of future DV screening projects, with
particular attention to the impact of the availability of intervention on screening and
disclosure.
Limitations

One weakness of this study is that it continued the trend of focusing on lower-
income populations in research. It seems that the privacy of these individuals is often
seen as less valuable than that of higher-income groups, not only in research, but in
broader areas of our society as well. Not only does it seem unjust to continue intruding
for the purpose of research, but this pattern may also affect the results of a project such as
this one. Because of the greater intrusiveness into these women's lives, they may either be
more or less willing to disclose DV. These women are accustomed to being asked
personal questions and may not feel as much of a right to privacy as a result.
Alternatively, in response to feeling disrespected, they may be less likely to disclose such
personal and painful information to a health care provider. This setting was not chosen
for this project because of the SES of the patient population, but rather because it is the
site of the residency program and allowed for more provider training. This study needs to
be replicated in a private practice setting with higher-income women in order to

determine the effects of classism on the issue of DV screening.

I 104

A second weakness in evaluating the physician screening was that direct
observation was not done. Although all physicians attended the training workshop, it is
not known if their approaches to DV screening were similar. Different methods of
screening may have led to a differential rate of disclosure. Specifically, the ways in which
the screening questions are asked are as important as the questions themselves. For
example, a provider who screens every patient, but does so in an abrasive and uncaring
manner, may obtain fewer disclosures of abuse than a provider who only screens
occasionally but is empathic and caring in his/her approach. While the subject of how to
conduct the screening was discussed with the physicians, none of them were observed
with patients to assess actual behavior. .

Additional limitations to this project were those that are common to community-
based research. For instance, some of the data collected were incomplete due to the
number of individuals responsible for data collection. The cooperation of all clinic staff
members was needed to ensure that the AAS form was placed in each chart, completed,
and available for the researcher to review. For example, based on the project training, it
was hoped that physicians would have screened more often than they did. This project
required that changes be made in clinic procedures, and although these changes were
relatively minor, change is often difficult to bring about. These difficulties were likely
heightened by the fact that the researcher was not on staff at the clinic, or an Ob/gyn
professional. The researcher entered the relationship from the outside, with no previous
interaction with clinic staff. Taking the time to build a long-term relationship prior to
implementing such an intervention often increases the likelihood that the intervention will

be carried out as planned and fewer problems will occur. The final results may have been

105

effected had a collaborative relationship been established much earlier; the AAS may
have been correctly placed in each chart and the screening rate may have been higher,
possibly changing the disclosure data that were obtained. A longer-term intervention is
also helpful in this way, since behavior change occurs slowly but often does improve over
time. Finally, community-based research does not allow for the same degree of control as
laboratory research. For example, changesin the residency program staff, vacations, and
scheduling conflicts, were not within the control of the researcher, and had to be worked
into the project plan as needed.

Implementation Issues

 

Several discussions between the principal investigator and the residents illustrated
some of the problems with implementing the screening procedures as planned. Logistical
issues, such as where the screening form was kept in the chart and if the form had even
been placed in the chart prior to the exam were common problems. Approximately
halfway through the project implementation period, residents reported not always having
the AAS readily available. At that time, the charts of all pregnant patients were pulled
and the AAS was placed in any chart in which it was missing.

The fiequency of screening was also stated as an issue; during the training,
screening at every prenatal visit was recommended, with the minimum screening being at
the first prenatal visit, at least once during each trimester, and at the postpartum visit.
Some residents reported that they disagreed with the need for such frequent screening and
were not screening at all as a result. It was agreed that four screenings (first visit in first
trimester, second trimester, third trimester, and postpartmn) would be sufficient, and

residents were urged to begin screening patients if they had not already done so. In

106

addition, some doctors reported their patients became angry if they were asked more than
once, after screening negative the first time. Various ways of responding to this situation
and how to ease a patient’s mistrust of repeated screenings were discussed.

Several doctors also indicated dissatisfaction with the AAS itself; it was too
structured and sounded unnatural, and it was too long. We discussed ways of asking the
questions in one’s own words and they agreed the more times they asked the questions
the easier it became. Some residents admitted having some discomfort in addressing the
issue of partner violence with their patients at all, but they agreed to use the AAS and try
to increase their comfort through practice. Support from the residency program directors
was an important part of the project’s success, as they reportedly sent residents back into
exam rooms to complete the screening and frequently checked to insure that screening
was occurring.

Finally, the clinic record keeping led to difficulties in several areas. As
mentioned, race/ethnicity was not routinely recorded and this information was difficult to
find in the medical charts. In addition, there were inconsistent records of the number of
prenatal visits in the charts, making it hard to correctly determine the number of visits in
relation to abuse disclosure. Finally, it was not possible to identify which patients were in
the MSSP program, due to unsystematic recording of this information. Repeated
discussions with clinic staff were helpful in understanding the chart system, but did not
rectify these problems that impeded data collection.

Future Directions

 

There is no question that screening for DV during pregnancy is important; the

question is how to minimize barriers in order to best implement effective screening

107

procedures. This study added to the available literature by applying a theoretical
framework to identify potential barriers, and evaluating how one method of screening
was implemented in a particular type of prenatal care setting.

The data presented here indicate that it does not make a difference which health
care provider screens for DV. This issue should be addressed in studies in different
settings and over a longer period of time, to determine if this finding remains consistent.
There is likely a benefit to being screened by the same person each time, whomever that
provider is, as part of building a trusting relationship. Although few patients in this study
disclosed after the first screening, it is still recommended that screening continue
throughout pregnancy. Providers should follow the ACOG recommendations, which are
to screen at routine Ob/Gyn and preconception visits, first prenatal visit, at least once
each trimester, and at the postpartum visit.

One area for future study is to replicate a similar project in a private practice
setting. It is necessary to determine if the barriers to screening and disclosure are similar,
before the results of this study can be generalized to other settings. It is possible that
some of the cultural issues examined here, with a primary focus on race/ethnicity, are
products of socioeconomic status and classism. There may be different interactions when
patient and doctor are considered more as peers, that would eliminate some barriers and
produce different barriers to effective screening.

As stated previously, a long-term intervention and evaluation is necessary to more
fully understand the process of DV screening. For example, a woman might see the same
physician through multiple pregnancies and continued screening and interactions between

doctor and patient might be examined. Furthermore, the structural changes that are

108

necessary to make DV screening more effective take time to implement, and the long-
term impact of making such changes needs to be analyzed. It would also be possible to
examine provider change over time, to determine if comfort and effectiveness in
screening continue to improve.

Future research can build on the theoretical model used in this study. One way to
advance this research is to modify The Barriers Model so it can be tested empirically.
Through such research, one would be able to quantify the impact of specific barriers, and
of each level of barriers. It might then be possible to more precisely determine which
barriers have a greater impact on the screening process and how to best minimize this
impact. A statistical model would also aid in understanding why and how these barriers
act on the DV screening process.

Finally, an improved health care response to domestic violence, including
screening and intervention, must be conducted within the context of a greater societal
response. It must coincide with stronger prevention efforts, as well as more effective
intervention with abuse victims in other areas, such as the criminal justice and mental
health systems. As long as abused women continue to be stigmatized and blamed for the
violence perpetrated against them, some will continue to believe that disclosure to
healthcare providers is ineffective. Healthcare providers, who are viewed with respect and
often have a strong voice in their communities, can serve an important role in changing

the response to DV at the societal level.

109

APPENDIX A

Provider Training Outline

110

I.

II.

III.

Introduction
A. Presenters’ background
1. project investigator
2. community education coordinator
B. Overview of session
C. Respect for people in room who have experienced violence

Overview of Basic DV Information (community education coordinator)
A. Statistics
1. general facts
2. prevalence
B. What is Domestic Violence?
1. Power and control wheel
2. Cycle of violence theory
C. Victims and perpetrators
1. causes of battering
2. barriers to leaving

DV and Cultural Issues
A. Definition of culture and cultural competency
1. Culture is the shared experiences or commonalities that groups of individuals
have developed in relation to changing social and political contexts.
2. Culture may be based on race, ethnicity, sexual orientation, gender, religion,
age, class, immigration status, disability status, or other axes of identification
a. Also includes things like family structure, gender roles, marriage patterns,
sexual behavior, contraceptive patterns, pregnancy and childbirth
practices, child rearing practices, housing arrangements, migrant status,
level of acculturation into dominant culture, occupation, use of chemical
comforters, self-treatment strategies, lay therapies
b. It is not any one of these factors, but many of them in combination that
defines an individual’s culture.
3. Culture (many factors acting in combination) will define each woman’s
subjective experience of being abused
4. There is diversity within cultures as between cultures, and a patient’s beliefs
may be affected by the influence of culture at many levels
5. Both the provider’s and the patient’s cultures impact the interaction between
them
a. example of minority woman being hesitant to disclose abuse to Caucasian
doctor; any woman disclosing to male provider
b. example of hesitancy to screen white, middle class women for DV
6. Define cultural competency and relate it to working with DV issues
a. three elements of the process
b. three ways of being culturally competent when working with DV victims

lll

B. Need for cultural sensitivity in the screening process

1. How culture can serve as a barrier to disclosure

2. Awareness of one’s own culture being a barrier to screening
C. Interaction of culture and DV

(effects at intrapersonal, interpersonal, and structural levels)

1. Racial/ethnic minorities
stigmatization of group by disclosing DV
cultures which value family or kinship over individual may lead to intense
shame (seen in higher rates of suicide in non-white ethnic groups)
African American women (many issues are relevant to other women of
color as well)

a.
b.

1. greater flexibility between African gender roles may lead to conflict
with “American” gender roles

2. African-American women may be more likely to disclose DV to
African-American provider

3. leaving the relationship is often not the primary goal (especially in
low-income families) and providers should not impose their beliefs on
the patient -- economic survival may be greater priority and these
women may need referrals for services other than DV shelter and
counseling

4. often rely on friends or relatives as primary support system, but may
turn to health care providers if their support system is not available or
if violence is severe

Hispanic/Latina women

1. must consider differences among subgroups and level of acculturation

2. extended kinship system, rigid gender roles, large power differential
between genders

3. Mexican American women have reported greater rates of marital rape
and emotional abuse than non-Hispanic women in some studies

4. often have to consider other cultural issues, such as economic
situation, language, immigration status, religion

5. rarely seek services outside of community (in research literature)

Asian women - also great variety within Asian culture

1. family more important than individual

2. power differential, lack of access to resources, pressure to maintain
harmony in marriage, sexism within culture all contribute to DV in this
community

3. reporting of abuse by an Asian woman may indicate that it is quite
severe

4. extreme shame and fear of being ostracized

5. Indian culture still sees dowry-related deaths and “honor killings”;

there is research evidence of increased DV during pregnancy in Indian
women

Immigrant or refugee women

1.

women may be unaware of immigration laws and fear deportation

112

 

2. often poor, uneducated, may not speak English, do not have legal
residency, lack of social support, stress of immigrating
Do NOT use family members as interpreters when screening for DV if
the patient is unable to speak English
3. Class - issues of poverty may outweigh DV and decrease chance that she will
seek “professional” services

4. Gender

5. Patient vs. Provider

6. Religion

7. Sexual orientation (Lesbian battering)
8. Patients with disabilities

D. Recognize that racism, sexism, classism, and other forms of oppression might
impact a woman’s ability to cope with an abusive relationship, as well as her
available resources (Some DV intervention and community resources may not be
sensitive to cultural differences and may lead to patient’s reluctance to seek
services from these agencies)

Respect for patient’s diversity and sensitivity to their experiences will inevitably
reduce victim-blaming attitudes and enhance providers’ ability to work with women
who are victims/survivors

IV. DV and Women’s Health
A. General (potential) health effects
1. Physical injuries and warning signs
a. pain: headaches, chest pain, back pain, pelvic or abdominal pain
b. exacerbation or poor control of chronic medical conditions
c. serious illness, e.g. HIV
(1. perpetrator behavior as a warning sign
2. Mental health effects
depression
anxiety and panic disorders, PTSD
substance use/abuse
psychotic symptoms
eating disorders
somatoform disorders
g. suicide
B. DV during pregnancy
1. prevalence of 0.9% to 20.1%
a. most studies between 3.9% and 8.3%
b. rates were higher (16% to 20.1%) when screening occurred more than
once, in person, and continued throughout pregnancy
c. rates are much higher for verbal or psychological abuse
d. some studies have found reported rates of abuse increase in third trimester
and postpartum period

rape-99‘s»

113

e. prevalence is higher for adolescents (43.7%), who are at risk for abuse
from partner and family members

f. more common pregnancy complication than hypertension, preeclampsia,
hyperthyroidism, or gestational diabetes (which are routinely screened for)

. why battering during pregnancy

a. abuse related to woman’s increased vulnerability and defenselessness

b. anger and/or jealousy towards fetus, who may be seen as competition or as
interfering with abuser’s control

c. stress of family transition, including economic pressures and sexual
“fi'ustration”

(1. “business as usual” - continuation of previous abuse

. increased risk vs. protection of pregnancy

a. About one-half of women who are battered before pregnancy are also
battered during pregnancy

b. for some women who are abused prior to pregnancy, this may be the only
time the abuse subsides, or it may be that she is no longer with that parmer

. specific health effects to woman and fetus

a. direct result of physical tramna

l.
2.

3.

abdomen and reproductive organs are often targeted

abdominal trauma can cause fetal injury or death, with trauma to the
uterus sometimes resulting in miscarriage

usually not one incident, but repetitive trauma and trauma that occurs
in third trimester and involves direct blows to the abdomen, that leads
to fetal demise

b. preterm labor, chorioamnionitis, fetal distress or demise, infants
hospitalized longer, placental separation, antepartum hemorrhage, fetal
fracture, uterine rupture

c. sexual abuse

1.

2.
3.

one-fourth to over half of physically abused women are also sexually
abused '

pregnancy may be a result of rape

pregnancy may be because of inability to control contraceptive use,
alcohol use, early sexual initiation, or unprotected sex

SA increases risk for miscarriage or stillbirth, complicated deliveries,
breast-feeding problems, endometriosis, poor healing of episiotomies
or lacerations, STD’s (associated with premature membrane rupture,
preterm birth, chorioamnionitis, endometriosis, genital and peritoneal
infections, infertility, ectopic pregnancy), dyspareunia, chronic pelvic
pain, sexual dysfunction, frequent vaginal and urinary tract infections

(1. indirect effects

1. DV is associated with late entry to care and inconsistent check-ups
a. late ETC is a risk factor for LBW and other risks to
fetal and maternal health
b. batterer may feel threatened by her developing a
relationship with her doctor

114

c. women may miss appointments to avoid having
physical signs of abuse seen

2. battered women are 2-4 times more likely to give birth to LBW
infants

3. stress may affect woman’s self-care during pregnancy, often
related to smoking and substance abuse (which are related to
adverse birth outcomes), unhealthy diet, poor weight gain,
homelessness following separation from partner, tension and
arguments with partner, more worrying, negative attitudes about

pregnancy

e. maternal mortality

1. homicide and suicide are significant cause of maternal mortality
cross—culturally

2. more women are killed by abusive partners than any single medical
complication

3. homicide is leading cause of injury-related death among pregnant
women in US. urban areas

C. Improved health care response
attitudes toward battered women in the health care setting
2. prevention/intervention

1.

a.
b.
c.

expose abuse of individual women and provide intervention
expose issue of IPV to public and health care community
increase safety for women

3. FVPF’s elements of improved response

9.09:9:

routine screening
assessment
intervention
documentation

V. DV Screening

A. Screening in the health care setting
J CAHO and AMA guidelines require comprehensive identification,
documentation, treatment, and referral procedures
Upcoming ACOG recommendations
Patient wishes versus provider practices (Question audience first on how they
think patients feel)

1.

2.
3.

a.

b.

lack of DV communication can lead to dissatisfaction of care, more
trouble communicating in general *

being blamed or ignored by provider leads patients to see providers as
ineffective sources of help

most female patients favor routine assessment of physical and sexual
assault, and would report honestly if asked directly by a health care
provider; abused women are often relieved that someone asks, and
nonabused women often know someone who is abused and would like
information about DV

115

(1. patients may not volunteer information, but will answer honestly if asked

6.

B. Tips

directly; they believe that their providers should initiate the discussion
Findings from ACOG survey: most aware of DV issues, felt comfortable
about screening, and believe that patients are not adverse to being
questioned, but most (68%) only screen when abuse is suspected and less
than 5% screen after first trimester

l. screening
a.

b.

atoms»

H. p...
O O

k.

1.

must be done in private, separate from partner (decide what to do about
having children present)
establish rapport prior to screening

c. assure patient of confidentiality
(I.

present screening as routine, something asked of all patients because of the
prevalence of the problem

ask direct questions

be calm, matter-of-fact, and non-j udgmental

listen to patient, allowing her to tell her perspective

screen at initial prenatal visit and all subsequent visits, including
postpartum

when needed, use professional interpreters, not family or fiiends of patient
will take less time than looking for risk factors first and only asking when
abuse is “suspected”

educates patients by showing that IPV is a significant health issue and
health care providers can be used as a resource

ask about abuse like you would ask about anything else

2. assessment
a. follow up to get more details, pattern and history of abuse
b. assess current danger to patient
3. intervention tips
a. don’t need to feel responsible for all aspects of screening and intervention;

work with other providers and other agencies to make referrals and have a
more coordinated response ‘

b. mention safety planning and documentation
4. making referrals

a. ask and listen to what the patient needs and wants
b. express concern and support for the patient
c. alert the patient to any confidentiality issues or reporting requirements
(I. know your local services and provide a list of services to patients so that
they can select the most appropriate services
e. ask patients if they want help making the contact
f. remind patients that they can speak to you again about the violence
g. make referral to clinic social workers
C. Reporting

1. State reporting requirements (full text in resource manual, along with M1 DV
related statutes from MCADSV Resource Center)

116

 

2. How reporting can negatively affect patient-provider relationship
patient at risk for abuser retaliation (
deterring patient from seeking health care or being open about abuse
cause provider to pass responsibility for patient care to another agency
harming patient if her safety is not accounted for by responding system
infiinging on patient’s autonomy and right to make her own decisions
disrupting confidentiality, which may lead to decreased trust in provider
and open relationship with provider (including discussions of options and
resources
D. Patient barriers to disclosure
1. fear of batterer, fear of safety or children’s safety, lack of economic resources,
previous experience with criminal justice or social service systems, cultural
barriers such as language and inunigration factors
2. values/beliefs about relationships, religion, family of origin values
impaired ability to see danger, reduced self-esteem, brainwashing, isolation,
thinking she deserves the abuse/does not deserve help, minimizing the abuse
depression, substance abuse, anxiety, PTS responses from long-term abuse
shame, humiliation
lack of trust in provider (e. g. may report to police, lack of confidentiality)
provider discomfort with topic
power differential (medical culture of provider vs. patient)
9. protecting partner (still loves him, be financially supports family)
10. not realizing the harm the stress of being abused has on her physically
E. Provider barriers to screening
1. Review lists of potential barriers from literature, have audience add to list
2. Refuting myths that support some of the barriers (e.g. time, offending
patients)
3. What you can and cannot do for the patient

W999?!”

S”

FM“???

VI. Training on AAS
A. Purpose of each question
B. Possible “lead-ins” to screening questions
C. Modeling
D. Be aware of own responses to hearing women’s stories and be sure that you have
some sort of support network to deal with the impact

VII. Community Resources
A. Resource manual

ll7

APPENDIX B

Patient Survey

118

 

MSU DV STUDY

You are being asked to participate in a study of patients’ views on domestic violence and
how patients feel about sharing personal information with their health care providers. We
are giving this survey to all women, age 18 or over, who come to the Ob/Gyn clinic for
any type of appointment. Please do not complete this survey if you are under 18. Please
take a few minutes to answer these questions honestly. Participation is completely
voluntary, and you can choose not to participate at all, or not to answer certain questions
if you choose, with no penalty. This survey is completely anonymous - your name will
not appear anywhere on this survey. You indicate your voluntary agreement to participate
by completing and returning this questionnaire. If you choose to participate, please
complete the questionnaire now and return it to the person that gave it to you.

If you have any questions or concerns about this study, please contact project investigator
Deborah Shapiro at 353-9858. For questions about participants’ rights as human subjects
of research, please contact UCRIHS Chair David Wright at 355-2180 or Sparrow IRRC
Chair George Abela at 483-2164.

119

 

1. What is your age? years

2. Are you (check one): single married divorced
separated widowed

3. Do you consider yourself (please check one):

 

 

 

White/Caucasian Black or African American Native American
Asian or Pacific Islander Hispanic/Latina Middle Eastern
Biracial (please specify: )
Multiracial (please specify: )
Other (please specify: )
4. What is the highest grade in school you have completed? (check one)
Less than high school High school graduate GED
Some college College graduate Advanced degree
5. Are you currently pregnant? YES NO
If yes, what trimester are you in? lst 2nd 3rd
6. Approximately how many prenatal appointments have you had? visits

7. Please circle the letter of each item you think best describes the term Domestic Violence
(circle all that apply):

Pushing, shoving, slapping, punching, kicking, choking

Assault with a weapon

Holding, tying down, or restraining you

Threats of harm, or intimidating you

Keeping you from seeing your friends or family

Degrading, humiliating, or making you feel bad about yourself

Trying to make you perform sexual acts against your will

Hurting you physically during sex or assaulting your genitals

Forcing you to have sex without protection against pregnancy or sexually transmitted
diseases

r‘?q°r~9999'e

8. How comfortable would you be talking about violence/abuse with each of the following
people? (please circle one number for each statement below):

Very Somewhat Somewhat Very
Uncomfortable Uncomfortable Not sure Comfortable Comfortable

a. The nurse taking care of you 1 2 3 4 5

b. Your doctor 1 2 3 4 5

c. A social worker at the clinic 1 2 3 4 5

d. A counselor from the community 1 2 3 4 5
9. Has your doctor ever asked you if you have experienced domestic violence? _ YES _NO
10. If your doctor asked, would you tell him/her ifyou were being abused? YES NO

120

1 1. Please circle the number that shows how much you agree or disagree with each ofthe

following statements:

Strongly Strongly
Disagree Disagree Not sure Agree Agree
1 am comfortable discussing personal issues 1 2 3 4 5
with my doctor.
I am more comfortable discussing personal 1 2 3 4 5

issues (such as violence/abuse) with a female doctor.

I am more comfortable discussing personal I 2 3 4 5
issues (such as violence/abuse) with a male doctor.

I am more comfortable discussing personal 1 2 3 4 5
issues (such as violence/abuse) with a doctor who

is the same race/ethnicity as me.

I believe that domestic violence should not I 2 3 4 5
be discussed outside the family.

1 would tell my doctor ifl was being abused. l 2 3 4 5

1 would know where to go for help if my I 2 3 4 5
partner was abusing me.

I would be offended if my doctor asked me I 2 3 4 5
about domestic violence/abuse.

1 would be concerned about discussing abuse I 2 3 4 5
with my doctor because of how my partner would react.

1 do not believe that the police would treat me I 2 3 4 5
fairly if 1 reported domestic violence.

12. What are some reasons you might n_gt tell your doctor if you were being abused? (Please list
up to 5):

l.

2.

 

 

 

 

 

Thank you for taking the time to fill out this survey!
Please feel free to add any comments on the back.

APPENDIX C

Provider Pre-Training Survey

122

Survey on Domestic Violence

You are being asked to participate in a study of health care professionals’ views
concerning domestic violence. We are using the ACOG definition of abuse, which states
that domestic abuse is any act by a person who is currently, or was previously, in an
intimate relationship with the victim, that is intended or perceived to be intended, to
cause physical or psychological harm. Participation in this study is completely voluntary
and you may choose to not participate or to answer certain questions without penalty. You
indicate your voluntary agreement to participate by completing and returning this
questionnaire. This survey is confidential, but we ask that you enter the last four digits of
your phone number (or another four digit number that you will be sure to remember) in
the upper right-hand comer as a way of tracking the surveys. Please complete and return
the survey in the envelope provided by September 10’“. Thank you for taking the time
required to fill out this survey.

If you have any questions or concerns about this study, please contact project investigator
Deborah Shapiro at 353-9858. For questions about participants’ rights as human subjects
of research. please contact UCRIHS Chair David Wright at 355-2180 or Sparrow IRRC
Chair George Abela at 483-2164.

Clinical Practice
1. Which best describes yourjob position?

Nurse Practitioner Physician

Nurse Midwife Social Worker
Clinical Nurse Specialist Psychologist
Nurse Administrator Physician Assistant

Other (please specify)

 

2. Highest degree held:

3. Years working in an Obstetrical/Gynecological Clinic:
4. Current year of residency (if applicable):

5. Your age: __ years

6. Gender: __ Female ___Male

7. Do you consider yourself (please check all that apply):

 

 

Black or African American Native American

White , not Hispanic Asian or Pacific Islander

Hispanic/Latino Biracial (please specify: )
Multiracial (please specify: )

Other (please specify: )

 

8. Some providers ask female patients direct, specific questions about abuse while others do not. Do you
ask patients direct, specific questions about whether they have been abused or threatened? (please

circle one):
1 2 3 4 5
Never Only ifl suspect abuse At the initial visit Often At every visit

123

9. When you see female patients under the following circumstances, how frequently do you (or other staff
members) ask direct, specific questions about abuse? (Please circle one response for each statement

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

below):
.0 3
,_ 5: fi 33 E a
Z O to U) U) Q <
a. Initial visit 1 2 3 4 5
b. Annual exams/regular checkups 1 2 3 4 5
c. Patient presents with injury 1 2 3 4 5
d. Patient complaints/syndrome suspicious l 2 3 4 5
e. First prenatal visit 1 2 3 4 5
f. Sometime during lst trimester l 2 3 4 5
g. Sometime during 2nd trimester l 2 3 4 5
h. Sometime during 3rd trimester l 2 3 4 5
1. At an antenatal visit 1 2 3 4 5
j. At a postpartum visit 1 2 3 4 5
k. At every visit 1 2 3 4 5
1. If the patient raises the subject first 1 2 3 4 5
10. Would you do any of the following if a patient of yours was identified as bein abused?
3
g g,
E E E E
2 (I) O <
a. Provide the patient with written information about community resources 1 2 3 4
b. Document the abuse in the patient’s chart 1 2 3 4
c. Include a body map in patient’s chart to document injuries 1 2 3 4
d. Photograph injuries 1 2 3 4
e. Report the abuse to the police or local authority 1 2 3 4
f. Ask about the patient’s safety 1 2 3 4
g. Ask about the safety of patient’s children 1 2 3 4
h. Ask whether there are weapons in patient’s home 1 2 3 4
i. Schedule a follow-up visit regardless of the patient’s health status I 2 3 4
j. Admit a patient to the hospital because you thought she was in danger 1 2 3 4
k. Testify in court about the abuse of your patient 1 2 3 4
1 1. Which of these other types of screening do you routinely perform as part of your practice?
(please check all that apply)
__ Inquire about past sexual assault (e.g. rape) _ Inquire about use/abuse of alcohol
__ Inquire about childhood physical/sexual abuse __ Inquire about use/abuse of other drugs
__ Inquire about stress or depression __ Inquire about use of seat belts
__ Inquire about use/abuse of tobacco __ Inquire about diet and exercise

12. Does your practice or institution have a written protocol for screening patients for domestic violence?
YES NO Don’t know

13. Do you conduct a one-on-one discussion with each patient as part of her exam?
YES NO Don’t know

14. Do you allow or encourage partners to accompany patients during exams?
YES NO Don’t know

124

Attitudes/Opinions
Please circle the response that best reflects your own level of agreement or disagreement with each of the

following statements:

20.

21.

22.

23.

24.

25.

26.

27.

28.

30.

 

 

 

 

 

 

 

125

.2» 8 8 95’ b

25’5“ a a §” 8
b .‘L’ .."..’ O 00 j: 31)
to Q D 2 <2 (I: <

. Domestic violence may begin or escalate during pregnancy. 1 2 3 5

. There is no single profile of a domestic violence victim. I 2 3 5

. Women are at an increased risk of being killed when they 1 2 3 5
decide to separate from their partner.

. The effects of emotional abuse can be longer lasting than 1 2 3 5
those of physical abuse.

. Domestic violence (DV) is an issue of power and control. 1 2 3 5
Intervention with victims of domestic violence who are l 2 3 5
unwilling to leave their abusive partner is ineffective.

DV can have an important impact on my patients’ health. 1 2 3 5
DV can have an important impact on my patients’ 1 2 3 5
pregnancy outcome.

Battered women are almost always from groups with l 2 3 5
low socioeconomic status.

Health care professionals can do little to help a woman 1 2 3 5
who is being abused.

There are sufficient resources in my community 1 2 3 5
to assist abuse victims.

I believe 1 can help a patient who is being abused l 2 3 5
by her partner.

I feel comfortable asking patients direct questions about abuse. I 2 3 5
Patients have a right to privacy about family matters like abuse. I 2 3 5

. Most female patients will deny abuse if asked. 1 2 3 5
DV is a way of life in many minority communities. 1 2 3 5

 

3l.

32.

33.

34.

35.

36.

37.

 

 

 

 

 

 

 

2» 8 8 g. 2»
on 5'!) |51) 0 °0 0
8 a :9, .‘9 2 8 £30
b .. .... 0 60 h:
m D D Z < m <
Health care providers have as much responsibility to deal l 2 3 5
with partner abuse as they do to'deal with other clinical problems.
I have a lack of knowledge about domestic violence 1 2 3 5
resources in my community.
I am concerned that ifl ask about DV, I will I 2 3 5
offend my patients.
I am concerned that if I ask about domestic abuse. I 2 3 5
I may inflame an abusive situation. '
I received an adequate amount of training and information I 2 3 5
on domestic violence during my degree training program or in subsequent training.
I think there should be more DV education and training I 2 3 5

for health care professionals.

How confident do you feel about your skills in asking patients about domestic violence? (please circle

oney

 

 

I 2 3 4
Not very confident Unsure Somewhat confident Very confident

. There are many obstacles to asking patients about abuse. In your own practice, please indicate how
important each of the following factors are to your decision whether to ask about abuse. (please circle
your response):

Not Slightly Quite Very
Important Somewhat Important
a. Time constraints make it difficult to ask about abuse. I 2 3 4 5
b. It is frustrating to identify abuse because I can
do little to help. I 2 3 4 5
c. Direct questioning about abuse may be too
confrontational for my patients. I 2 3 4 5
d. My patients would be embarrassed if I asked them I 2 3 4 5
about abuse.
e. Language barriers make it difficult to discuss
abuse with patients. I 2 3 4 5
f. Different cultural beliefs and values make it
difficult to discuss abuse with patients. I 2 3 4 5
g. My attention is focused on other health
problems that are of higher priority. I 2 3 4 5

126

h. I am uncomfortable discussing DV with my patients. I 2 3 4 5

i. Other, please specify: 1 2 3 4 5

 

THANK YOU FOR TAKING THE TIME TO COMPLETE THIS SURVEY

Please feel free to write any additional comments on the back.

127

APPENDIX D

Provider Post-Training Survey

128

Survey on Domestic Violence

You are being asked to participate in a study of health care professionals’ views concerning
domestic violence. We are using the ACOG definition of abuse, which states that domestic abuse
is any act by a person who is currently, or was previously, in an intimate relationship with the
victim, that is intended or perceived to be intended, to cause physical or psychological harm.
Completion of this survey is voluntary and you may choose to not respond or answer certain
questions without penalty. You indicate your voluntary agreement to participate by completing
and returning this questionnaire. This survey is completely confidential. Thank you for taking the
time required to complete this survey.

If you have any questions or concerns about this study, please contact project investigator
Deborah Shapiro at 353-9858. For questions about participants’ rights as human subjects of
research, please contact UCRIHS Chair David Wright at 355-2180 or Sparrow IRRC Chair
George Abela at 483-2164.

Clinical Practice

1. Which best describes yourjob position?
Registered Nurse Physician Medical Assistant
Nurse (other) Social Worker

2. Years working in an Obstetrical/Gynecological Clinic:

3. Your Age (in years): years

4. Your Gender: Female Male

5. Do you consider yourself (please check all that apply):

 

 

Black or African American White, not Hispanic Hispanic/Latino
Asian or Pacific Islander Native American

Biracial (please specify: )

Multiracial (please specify: )

Other (please specify: )

 

6. Some providers ask female patients direct, specific questions about abuse while others do not. Do you
ask patients direct, specific questions about whether they have been abused or threatened? (please

circle one):
I 2 3 4 5
Never Only if I suspect abuse At the initial visit Often At every visit

7. How confident do you feel about your skills in asking patients about domestic violence? (please circle
oney
I 2 3 4
Not very confident Unsure Somewhat confident Very confident

8. Does your practice or institution have a written protocol for screening patients for domestic violence?
YES NO Don’t know

9. Do you conduct a one-on-one discussion with each patient as part of her exam?
YES NO Don’t know

10. Do you allow or encourage partners to accompany patients during exams?
YES NO Don’t know

129

 

l I. When you see female patients under the following circumstances, how frequently do you (or other staff
members) ask direct, specific questions about abuse? (Please circle one response for each statement

 

 

 

 

 

 

 

 

below):

t... .9 B .E U)

3'5 ; g 8 5 c: g‘

S E g g. E E 2

Z O :6 m (I) O <

a. Initial visit I 2 3 4 5
b. Annual exams/regular checkups I 2 3 4 5
c. Patient presents with injury I 2 3 4 5
d. Patient complaints/syndrome suspicious 1 2 3 4 5
e. First prenatal visit I 2 3 4 5
f. Sometime during lst trimester 1 2 3 4 5
g. Sometime during 2nd trimester I 2 3 4 5
h. Sometime during 3rd trimester l 2 3 4 5
i. At an antenatal visit 1 2 3 4 5
j. At a postpartum visit I 2 3 4 5
k. At every visit I 2 3 4 5
I. If the patient raises the subject first I 2 3 4 5

Attitudes/Opinions
Please circle the response that best reflects your own level of agreement or disagreement with each of the
following statements:

 

 

 

 

 

 

 

 

b 8 8 “5’ .2:
is" in so i”. 8 is” 3
j: ‘2 .5 o '51) 1: 51)
m D D Z < m <
12. Women are at an increased risk of being killed when 1 2 3 4 5
they decide to separate from their partner.
13. Intervention with victims of domestic violence who are I 2 3 4 5
unwilling to leave their abusive partner is ineffective.
l4. Battered women are almost always from groups with 1 2 3 4 5
low socioeconomic status.
15. Health care professionals can do little to help a woman I 2 3 4 5
who is being abused.
I6. I believe I can help a patient who is being abused I 2 3 4 5
by her partner.
17. I feel comfortable asking patients direct questions I 2 3 4 5
about abuse.
18. Patients have a right to privacy about family matters 1 2 3 4 5
like abuse.
19. Most female patients will deny abuse if asked. 1 2 3 4 5

I30

20.

21.

23.

24.

25.

26.

27.

DV is a way of life in many minority communities.

Health care providers have as much responsibility
to deal with partner abuse as they do to deal with other
clinical problems.

. I have a lack of knowledge about domestic violence

resources in my community.

I am concerned that if I ask about DV, 1 will offend
my patients.

I am concerned that if I ask about domestic abuse,
I may inflame an abusive situation.

I received an adequate amount of training and
information on domestic violence during my degree
training program or in subsequent training.

I think there should be more DV education and
training for health care professionals.

1 found the DV educational session conducted
this year to be helpful.

. I find that I am increasingly comfortable asking

patients about relationship violence.

131

 

 

 

 

 

 

 

.233 8 95’ .29
é”? 5“ <2 §°8
13.3 .‘L’ 0 00 13a
:00 D Z < m<
l 2 3 4 5
l 2 3 4 5
l 2 3 4 5
I 2 3 4 5
l 2 3 4 5
l 2 3 4 5
I 2 3 4 5
1 2 3 4 5
l 2 3 4 5

 

29. There are many obstacles to asking patients about abuse. In your own practice, please indicate how
important each of the following factors are to your decision whether to ask about abuse. (please circle
your response):

Not Slightly Quite Very
Important Somewhat Important

i. Time constraints make it difficult to 1 2 3 4 5
ask about abuse.

j. It is frustrating to identify abuse because I can I 2 3 4 5
do little to help.

k. Direct questioning about abuse may be too I 2 3 4 5
confrontational for my patients.

I. My patients would be embarrassed if I l 2 3 4 5
asked them about abuse.

m. I feel that I need to develop a relationship 1 2 3 4 5
with my patients before asking them about abuse.

n. Language barriers make it difficult to discuss I 2 3 4 5
abuse with patients.

0. It is awkward to ask a patient’s partner to leave 1 2 3 4 5
the room in order to ask her about abuse.

p. Different cultural beliefs and values make it 1 2 3 4 5
difficult to discuss abuse with patients.

q. My attention is focused on other health I 2 3 4 5
problems that are of higher priority.

r. I am uncomfortable discussing DV I 2 3 4 5

with my patients.

30. Please list some additional reasons why you have found it difficult to screen all patients for domestic
violence. or barriers that you or others have faced in attempting to do so:

a.

 

b.

 

 

 

 

132

31. Please list any suggestions you may have for incorporating effective domestic violence screening into
prenatal care:

a.

 

b.

 

 

 

 

THANK YOU FOR TAKING THE TIME TO COMPLETE THIS SURVEY

Please feel free to write any additional comments.

133

 

APPENDIX E

Abuse Assessment Screen

134

 

Abuse Assessment Screen

Patient Ethnicity: L} Caucasian
.3 African American [3 Native American
1? Asian/Pacific Islander

B Hispanic/Latina

 

Cl Biracial (specify:
[3 Multiracial (specify:
1] Arab/Middle Eastern

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L1 Other (specify:

TODAY’S DATE:
1. Are you afraid of your No i, No D No Ll No .3 No D No 13 No 3 No
partner or anyone living in ,1 Yes [i Yes L1 Yes {:1 Yes I; Yes D Yes "J Yes E Yes
your home?
2. Have you repeatedly been , No No F No _‘ No . No T No 3 No L: No
called names. told you were ‘ Yes Yes " Yes Yes ‘ Yes " Yes " Yes Yes
worthless. ugly. or verbally
threatened by a partner or
someone important to you?
Has this happened since you No No ‘, . No No No . No . No J No
became pregnant? Yes Yes Yes Yes ’ Yes Yes 1 Yes i ‘ Yes
When was the last time this
happened?
3. Have you ever been hit, ;; No ‘No D No I J No 1) No L.) No ‘ 'No 1-: No
slapped. kicked or otherwise Yes 3 Yes D Yes [I Yes C} Yes ;: Yes a Yes 'J Yes
physically hurt by a partner
or someone important to
you?
Has this happened since you ,, No ,3 No C No . I No C No No L; No I.“ No
became pregnant? Yes Yes Yes 13 Yes ’3 Yes L, Yes Yes Yes
When was the last time this
happened?
4. Has anyone ever forced No No ‘ No No 1 No No ‘“ No -No
you to d0 sexual things that Yes . Yes ; Yes i." Yes Yes ._i Yes 1.: Yes 1.1 Yes
you did not want to do?
Has this happened since you . No No . .No . No No 3_ No No ' No
became pregnant? Yes Yes " Yes ’ , Yes “ Yes 7 Yes . Yes '1 Yes
When was the last time this
happened?
5. Who committed the abuse
(cg. ex or current partner)?
6. When did the abuse start?
Provider Initials:
Was a referral made? .. No . No .1 No . No . No . No .. No . No

Yes L Yes 1;: Yes 3 . Yes i 1‘ Yes II. Yes '; Yes [I Yes

If yes. to whom?

 

 

 

 

 

 

 

 

 

135

 

REFERENCES

Aciemo, R., Resnick, H.S., & Kilpatrick, D.G. (1997). Health impact of
interpersonal violence 1: Prevalence rates, case identification, and risk factors for sexual

assault. physical assault, and domestic violence in men and women. Behavioral Medicine,
2_3_. 53-64.

American College of Obstetricians and Gynecologists. (1995). ACOG technical
bulletin: Domestic Violence. International Journal of Gynecology & Obstetrics, 51, 161-
170.

Bachman, R., & Saltzman, LE. (1995). Violence against women: Estimates from
the redesigned survey (NCJ-154348). Washington, DC: US. Department of Justice.

 

Bash, K.L., & Jones, F. (1994). Domestic violence in America. North Carolina
Medical Journal, 55(9), 400-403.

Berenson, A.B., San Miguel, V.V., & Wilkinson, GS. (1992). Prevalence of
physical and sexual assault in pregnant adolescents. Journal of Adolescent Health, 13(6),
466-469.

Berenson, A.B., Stiglich, N.J., Wilkinson, G.S., & Anderson, GD. (1991). Drug
abuse and other risk factors for physical abuse in pregnancy among white non-Hispanic,
black, and Hispanic women. American Journal of Obstetrics & GynecologyLl 64(6),
1491-1499.

Berenson, A.B., Wiemann, C.M., Wilkinson, G.S., Jones, W.A., & Anderson,
GD. (1994). Perinatal morbidity associated with violence experienced by pregnant
women. American Journal of Obstetrics & Gynecology, 170(6), 1760-1769.

Bohn, D.K. (1990). Domestic violence and pregnancy: Implications for practice.
Journal of Nurse-Midwifery, 35(2), 86-98.

Bohn, D.K., & Holz, K.A. (1996). Sequelae of abuse: Health effects of childhood
sexual abuse, domestic battering, and rape. Journal of Nurse-Midwifery, 41(6), 442-456.

Bohn, D.K., & Parker, B. (1993). Domestic violence and pregnancy: Health
effects and implications for nursing practice. In Nursing Care of Survivors of Family
Violence (J. Campbell and J. Humphreys, Eds.). St. Louis, MO: Mosby-Year Book, Inc.
pp. 156-172.

Bullock, L.F., & McFarlane, J. (1989). The birth-weight/battering connection.
American Journal of Nursing, 89(9), 1153-1155.

136

 

Campbell, J .C., & Campbell, D.W. (1996). Cultural competence in the care of
abused women. Journal of Nurse-Midwifery, 41(6), 457-462.

Campbell, J .C., Harris, M.J., & Lee, R.K. (1995). Violence research: An
overview. Scholarly Inquiry for Nursing Practice: An International Journal, 9(2), 105-126.

Campbell, J .C., Oliver, C., & Bullock, L. (1993). Why battering during
pregnancy? AWHONN’s Clinical Issues in Perinatal and Women’s Health, 4(3), 343-349.

Campbell, J .C., Poland, M.L., Waller, J .B., & Ager, J. (1992). Correlates of
battering during pregnancy. Research in Nursing & Health, 15, 219-226.

 

Cohen, S., De Vos, E., & Newberger, E. (1997). Barriers to physician
identification and treatment of family violence: Lessons from five communities.
Academic Medicine, 72(1), 819—825.

Cokkinides, V.E., & Coker, AL. (1998). Experiencing physical violence during
pregnancy: Prevalence and correlates. Family & Community Health, 20(4), 19-37.

Coley, S.M., & Beckett, J .O. (1988). Black battered women: Practice issues.
Social Casework: The Journal of Contemporary Social Work, 69(8), 483-490.

Crawford, D.W., & Godbey, G. (1987). Reconceptualizing barriers to family
leisure. Leisure Sciences, 9(2), 119-127.

Crawford, D.W., Jackson, E.L., & Godbey, G. (1991). A hierarchical model of
leisure constraints. Leisure Sciences, 13(4), 309-320.

Crites, L. (1990). Cross-cultural counseling in wife beating cases. Response to the
victimization of women and children, 13(4), 8-12.

Dannenberg, A.L., Carter, D.M., Lawson, H.W., Ashton, D.M., Dorfman, S.F., &
Graham, EH. (1995). Obstetrics: homicide and other injuries as causes of maternal death
in New York City, 1987 through 1991. American Journal of Obstetrics and Gynecology.
1_'_7_2_(5), 1557-1564.

Dasgupta, S.D., & Warrier, S. (1996). In the footsteps of “Arundhatai”: Asian
Indian women’s experience of domestic violence in the United States. Violence Against
Women 2(3), 238-259.

 

Domestic Violence Project, Inc. (1994). The assessment and treatment of victims
of domestic abuse: model protocol. Wisconsin Medical Journal, 93(4), 178-182.

137

Dye, T.D., Tolliver, N.J., Lee, R.V., & Kenney, CJ. (1995). Violence, pregnancy
and birth outcome in Appalachia. Pediatric and Perinatal Epidemiology, 9, 35-47.

F ischbach, R.L., & Herbert, B. (1997). Domestic violence and mental health:
Correlates and conundrums within and across cultures. Social Science and Medicine,
4_5(8), 1161-1176.

 

Flitcraft. A. (1993). Physicians and domestic violence: Challenges for prevention.
Health Affairs. 12(4), 154-161.

 

Friedman, L.S., Samet, J .H., Roberts, M.S., Hudlin, M., & Hans, P. (1992).
Inquiry about victimization experiences: A survey of patient preferences and physician
practices. Archives of Internal Medicine, 152, 1186-1190.

Gagan, M.J. (I998). Correlates of nurse practitioners’ diagnostic and intervention
performance for domestic violence. Western Journal of Nursinflesearch, 20(5), 536-
553.

Ganley, AL. (1998). Improvingthe health care response to domestic violence: A
trainer’s manual for health care providers. San Francisco, CA: The Family Violence
Prevention Fund.

Gazmararian, J.A., Adams, M.M, Saltzman, L.E., Johnson, C.H., Bruce, C.,
Marks, .1.S., Zahniser, S.C., & The PRAMS Working Group. (1995). The relationship

between pregnancy intendedness and physical violence in mothers of newborns.
Obstetrics & Gynecology, 85(6), 1031-1038.

Gazmararian, J .A., Lazorick, S., Spitz, A.M, Ballard, T.J., Saltzman, L.E., &
Marks, J .S. (1996). Prevalence of violence against pregnant women. Journal of the
American Medical Association, 275(24), 1915-

Gerbert, B., Abercrombie, P., Caspers, N., Love, C., & Bronstone, A. (1999).
How health care providers help battered women: The survivor’s perspective. Women &
Health 29(3), 115-135.

 

Gerbert, B., Caspers, N., Bronstone, A., Moe, J ., & Abercrombie, P. (1999). A
qualitative analysis of how physicians with expertise in domestic violence approach the
identification of victims. Annals of Internal Medicine, 131(8), 578-584.

Gerbert, B., Johnston, K., Caspers, N., Bleeker, T., Woods, A., & Rosenbaum, A.

(1996). Experiences of battered women in health care settings: A qualitative study.
Women & Health, 24(3), 1-17.

138

 

Gielen, A.C., O’Campo, P.J., Faden, R.R., Kass, N.E., & Xue, X. (1994).
Interpersonal conflict and physical violence during the childbearing year. Social Science
& Medicine, 39(6), 781-787.

Gremillion, D.H., & Kanof, ER (1996). Overcoming barriers to physician
involvement in identifying and referring victims of domestic violence. Annals of
Emergency Medicine, 27(6), 769-773.

Grigsby, N., & Hartman, BR. (1997). The Barriers Model: An integrated strategy
for intervention with battered women. Psychotherapy, 34(4), 485-497.

Hayden, S.R., Barton, E.D., & Hayden, M. (1997). Domestic violence in the
emergency department: How do women prefer to disclose and discuss the issues? The
Journal of Emergency Medicine,15(4), 447-451.

Heise, LL. (1994). Gender-based violence and women’s reproductive health.
International Journal of Gynecology & Obstetrics, 46, 221-229.

Helton, A.S., & Snodgrass, F.G. (1987). Battering during pregnancy: Intervention
strategies. BIRTH 14(3), 142-147.

 

Heron, R.L., Jacobs, D.P., Twomey, H.B., & Kaslow, NJ. (1997). Culturally
competent interventions for abused and suicidal African American women.
Psychotherapy, 34(4), 410-424.

Horan, D.L., Chapin, J ., Klein, L., Schmidt, L.A., & Schulkin, J. (1998).
Domestic violence screening practices of obstetrician-gynecologists. Obstetrics &
Gynecology, 92(5), 785-789.

Huisman, K.A. (1996). Wife battering in Asian American communities:
Identifying the service needs of an overlooked segment of the US. population. Violence
Against Women, 2(3), 260-283.

Jang, D., Lee, D., & Morello-Frosch, R. (1990). Domestic violence in the
immigrant and refiJgee community: Responding to the needs of immigrant women.
Response to the Victimization of Women and Children, 13(4), 2-7.

Kantor, G.K., Jasinski, J .L., & Aldarondo, E. (1994). Sociocultural status and
incidence of marital violence in Hispanic families. Violence and Victims, 9(3), 207-222.

Kilpatrick, D.G., Resnick, H.S., & Aciemo, R. (1997). Health impact of
interpersonal violence 3: Implication for clinical practice and public policy. Behavioral
Medicine, 23, 79-85.

 

I39

 

King, MC. (1993). Changing women’s lives: The primary prevention of violence
against women. AWHONN’s Clinical Issues in Perinatal & Women’s Health, 4(3), 449-
457.

King, M.C., Torres, 8., Campbell, D., Ryan, J ., Sheridan, D., Ulrich, Y., &
McKenna, LS. (1993). Violence and abuse of women: A perinatal health care issue.
AWHONN’s Clinical Issues in Perinatal & Women’s Health, 4(2), 163-172.

Langford, DR. (1996). Policy issues for improving institutional response to
domestic violence. Journal of Nursing Administrators, 26(1), 39-45.

Liebschutz, J .M., & Rich, J .A. (1998). Shame, mistrust, and loss of control: A
qualitative analysis of barriers between battered women and physicians. Journal of
General Internal Medicine, 13(1), 108.

Lockhart, L., & White, B. (1989). Understanding marital violence in the Black
community. Journal of Interpersonal Violence, 4(4), 421-436.

McCauley, J ., Yurk, R.A., Jenckes, M.W., & Ford, DE. (1998). Inside “Pandora’s
box”: Abused women’s experiences with clinicians and health services. Journal of
General Internal Medicine, 13(8), 549-555.

 

McFarlane, J ., Christoffel, K., Bateman, L., Miller, V., & Bullock, L. (1991).
Assessing for abuse: Self-report versus nurse interview. Public Health Nursing, 8(4), 245-
250.

McFarlane, J., & Parker, B. (1994). Preventing abuse during pregnancy: An
assessment and intervention protocol. MCN/American Journal of Maternal and Child
Nursing, 19, 321-324.

McFarlane, J., Parker, 8., Soeken, K., & Bullock, L. (1992). Assessing for abuse
during pregnancy: Severity and frequency of injuries and associated entry into prenatal
care. Journal of the American Medical Association, 267(23), 3176-3178.

McGee, MP. (1997). Cultural values and domestic violence. Joumal of Famin
Social Work, 2(2), 129-140.

McGrath, M.E., Hogan, J .W., & Peipert, J .F . (1998). A prevalence survey of
abuse and screening for abuse in urgent care patients. Obstetrics & Gynecology, 91(4),
511-514.

Mayer, L., & Liebschutz, J. (1998). Domestic violence in the pregnant patient:

Obstetric and behavioral observations. Obstetrical and Gynecological Survey, 53(10),
627-635.

140

 

Mezey, G.C., & Bewley, S. (1997). Domestic violence and pregnancy. British
Journal of Obstetrics and Gynaecology, 104, 528-531.

Michigan Department of Community Health (1996). Survey of violence in the
lives of Michigan women. Lansing, MI: Author.

Molliconi, S.A., & Runyan, CW. (1996). Detecting domestic violence: A pilot
study of family practitioners. North Carolina Medical Journal, 57(3), 136-138.

Norton, L.B., Peipert, J .F ., Zierler, S., Lima, B., & Hume, L. (1995). Battering in
pregnancy: An assessment of two screening methods. Obstetrics & Gynecology, 85(3),
321-325.

Norton, I.M., & Schauer, J. (1997). A hospital-based domestic violence group.
Psychiatric Servicesflw), 1 186-1190.

O’Campo, P., Gielen, A.C., F aden, R.R., & Kass, N. (1994). Verbal abuse and
physical violence among a cohort of low-income pregnant women. Women’s Health
Issues 4(1), 29-37.

 

Olavarrieta, C.D., & Sotelo, J ., (1996). Letter from Mexico City: domestic
violence in Mexico. Journal of the American Medical Association, 275(24), 1937-1941.

Parsons, L.H., Zaccaro, D., Wells, B., & Stovall, T. (1995). Methods of and
attitudes toward screening obstetrics and gynecology patients for domestic violence.
American Journal of Obstetrics and Gynecology,l73(2), 381-386.

Petersen, R., Gazmararian, J .A., Spitz, A.M., Rowley, D.L., Goodwin, M.M.,
Saltzman, L.E., & Marks, J .S. (1997). Violence and adverse pregnancy outcomes: A
review of the literature and directions for future research. American Journal of Preventive

Medicine, 13(5). 366-373.

Phillips, D.S.H. (1998). Culture and systems of oppression in abused women’s
lives. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 27(6), 678-683.

Pinn, V.W., & Chunko, M.T. (1997). The diverse faces of violence: Minority
women and domestic abuse. Academic Medicine,72(1 supplement), S65-S7l.

 

Plichta, S.B., Duncan, M.M., & Plichta, L. (1996). Spouse abuse, patient-
physician communication, and patient satisfaction. American Journal of Preventive
Medicine, 12(5), 297-303.

Reichert, E. (1991). Perceptions of domestic violence against women: A cross-
cultural survey of international students. Response to the victimization of women and
children 14(1), 13-18.

 

141

 

Reid, A.J., Biringer, A., Carroll, J .D., Midmer, D., Wilson, L.M., Chalmers, B., &
Stewart, DE. (1998). Using the ALPHA form in practice to assess antenatal psychosocial
health. Canadian Medical Association Journal, 159(6), 677-684.

Rhee, S. (1997). Domestic violence in the Korean immigrant family. Journal of
Sociology and Social Welfare, 24(1), 63-77.

 

Ribe, J .K., Teggatz, J .R., & Harvey, GM. (1993). Blows to the maternal abdomen
causing fetal demise: Report of three cases and a review of the literature. Journal of
Forensic Sciences, 38(5), 1092-1096.

 

Rodriguez, M.A., Bauer, H.M., F lores-Ortiz, Y., & Szkupinski-Quiroga, S.
(1998). Factors affecting patient-physician communication for abused Latina and Asian
immigrant women. The Journal of Family Practice, 47(4), 309-311.

Rodriguez, M.A., Bauer, H.M., McLoughlin, E., & Grumbach, K. (1999).
Screening and intervention for intimate partner abuse: practices and attitudes of primary
care physicians. J AMA 282(5), 468-474.

 

Rodriguez, M.A., Szkupinski-Quiroga, S., & Bauer, H. (1996). Breaking the
silence: Battered women’s perspectives on medical care. Archives of Family Medicine,
5(3), 153-158.

Rosenberg, M.L, F enley, M.A., Johnson, D., & Short, L. (1997). Bridging
prevention and practice: Public health and family violence. Academic Medicine, 72(1),
S 1 3-818.

Santiago, A.M, & Morash, M. (1995). Strategies for serving Latina battered
women. In J .A. Garber & R.S. Turner (Eds), Gender in Urban Research (pp. 219-235).
Thousand Oaks, CA: Sage Publications, Inc.

Sheridan, D.J., & Taylor, W.K. (1993). Developing hospital-based domestic
violence programs, protocols, policies, and procedures. AWHONN’s Clinical Issues in
Perinatal & Women’s Health, 4(3), 471-482.

 

Sorenson, SB. (1996). Violence against women: Examining ethnic differences
and commonalities. Evaluation Review,20(2), 1230145.

Stewart, D.E., & Cecutti, A. (1993). Physical abuse in pregnancy. Canadian
Medical Association Journal, 149(9), 1257-1263.

Tjaden, P., & Thoennes, N. (1998). Prevalence, incidence, and consecmences of

violence against women: Findings from the national violence against women survey (NCJ
Publication No. 172837). Washington, DC: US. Department of Justice.

142

 

Ucko. LG. (1994). Culture and violence: The interaction of Africa and America.
Sex Roles, 31(3/4), 185-204.

 

Uzzell, O., & Peebles-Wilkins, W. (1989). Black spouse abuse: A focus on
relational factors and intervention strategies. The Western Journal of Black Studies,
13(1), 10-16.

Warshaw, C., & Ganley, A.L. (1995). Improving the health care response to
domestic violence: A resource manual for health care providers. San Francisco, CA:
Family Violence Prevention Fund.

Webster, J ., Sweett, S., & Stolz, T.A. (1994). Domestic violence in pregnancy: A
prevalence study. The Medical Journal of Australia, 161, 466-470.

Wiist, W.H., & McFarlane, J. (I 998). Utilization of police by abused pregnant
Hispanic women. Violence Against Women, 4(6), 677-693.

 

Yick, A.G., & Agbayani-Siewert, P. (1997). Perceptions of domestic violence in a
Chinese American community. Journal of Interpersonal Violence,12(6), 832-846.

 

143