454722
urnal of Interpersonal ViolenceHegarty et al.
© The Author(s) 2013
JIV28210.1177/0886260512454722Jo
Reprints and permission: http://www.
sagepub.com/journalsPermissions.nav
Article
Journal of Interpersonal Violence
28(2) 273–294
© The Author(s) 2013
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0886260512454722
http://jiv.sagepub.com
Effect of Type and
Severity of Intimate
Partner Violence on
Women’s Health and Service
Use: Findings From a Primary
Care Trial of Women Afraid
of Their Partners
Kelsey L. Hegarty, MBBS, FRACGP, DRANZCOG,
PhD,1 Lorna J. O’Doherty, BA, MBS, PhD,1 Patty
Chondros, BSc(Hons), GradDip Epidemiology &
Biostats, MSc(Statistics), PhD,1 Jodie Valpied, BA/
BTeach, PGDipPsych, COGE, MEd,1 Angela J. Taft, BA,
DipEd, MPH, PhD,2 Jill Astbury, BA, MEd, PhD,3
Stephanie J. Brown, BA(Hons), PhD,4 Lisa Gold, PhD,5
Ann Taket, BA, MSc, PhD,5 Gene S. Feder, BSc, MD,
FRCGP, DRCOG,6 and Jane M. Gunn, MBBS, FRACGP,
DRANZCOG, PhD1
Abstract
Intimate partner violence (IPV) has major affects on women’s wellbeing.
There has been limited investigation of the association between type and
severity of IPV and health outcomes. This article describes socio-demographic
characteristics, experiences of abuse, health, safety, and use of services in
women enrolled in the Women’s Evaluation of Abuse and Violence Care
1
The University of Melbourne, Melbourne, Australia
La Trobe University, Melbourne, Australia
3
Monash University, Melbourne, Australia
4
Murdoch Childrens Research Institute, Melbourne, Australia
5
Deakin University, Burwood, Australia
6
University of Bristol, Bristol, UK
2
Corresponding Author:
Kelsey L. Hegarty, MBBS, FRACGP, DRANZCOG, PhD, Associate Professor, General Practice,
University of Melbourne, 200 Berkeley Street, Melbourne,VIC 3053, Australia
Email: k.hegarty@unimelb.edu.au
274
Journal of Interpersonal Violence 28(2)
(WEAVE) project. We explored associations between type and severity of
abuse and women’s health, quality of life, and help seeking. Women (aged 16–
50 years) attending 52 Australian general practices, reporting fear of partners
in last 12 months were mailed a survey between June 2008 and May 2010.
Response rate was 70.5% (272/386). In the last 12 months, one third (33.0%)
experienced Severe Combined Abuse, 26.2% Physical and Emotional Abuse,
26.6% Emotional Abuse and/or Harassment only, 2.7% Physical Abuse only
and 12.4% scored negative on the Composite Abuse Scale. A total of 31.6%
of participants reported poor or fair health and 67.9% poor social support. In
the last year, one third had seen a psychologist (36.6%) or had 5 or more general practitioner visits (34.3%); 14.7% contacted IPV services; and 24.4% had
made a safety plan. Compared to other abuse groups, women with Severe
Combined Abuse had poor quality of life and mental health, despite using
more medications, counseling, and IPV services and were more likely to have
days out of role because of emotional issues. In summary, women who were
fearful of partners in the last year, have poor mental health and quality of life,
attend health care services frequently, and domestic violence services infrequently. Those women experiencing severe combined physical, emotional,
and sexual abuse have poorer quality of life and mental health than women
experiencing other abuse types. Health practitioners should take a history
of type and severity of abuse for women with mental health issues to assist
access to appropriate specialist support.
Keywords
domestic violence, anything related to domestic violence, assessment,
intervention/treatment
Intimate partner violence (IPV) is common in the community (Abramsky et al.,
2011; Australian Bureau of Statistics, 2005; Garcia-Moreno, Jansen, Ellsberg,
Heise, & Watts, 2006) and clinical populations, particularly in primary care
(Feder et al., 2009). It contributes to substantial physical and mental health
burden among women in developing and developed parts of the world
(Campbell, 2002; Ellsberg, Jansen, Heise, Watts, & Garcia-Moreno, 2008).
In Australia, IPV is the leading cause of morbidity and mortality for women
of child bearing age (Vos et al., 2006). Abused women use medical services
more frequently because of increased rates of mental health issues (depression, anxiety, suicide, somatization, posttraumatic stress disorder, substance
Hegarty et al.
275
abuse; Golding, 1999; Rees et al., 2011) and physical health issues (chronic
somatic complaints, reproductive problems and injuries; Campbell, 2002).
Despite the pervasive effects of IPV on societies, there are few trials of
interventions for IPV in health care settings. Furthermore, there is a paucity
of quantitative data on how different types and severity of abuse relate to
women’s health, quality of life, and service use, with most data coming from
the specialist service sector (e.g., Bonomi, Holt, Martin, & Thompson, 2006;
Feder et al., 2009; Straus et al., 2009; Wuest et al., 2010). Gathering this data
is hampered by most measures of IPV not characterizing type and severity of
abuse (Hegarty, 2006). In studies that have made these distinctions, higher
severity of abuse has been associated with lower mental and physical health
status, quality of life, and higher levels of depression, posttraumatic stress
disorder and chronic pain (Dutton, Kaltman, Goodman, Weinfurt, & Vankos,
2005; Straus et al., 2009; Wuest et al., 2010). The nature of these associations
has differed according to the type of abuse experienced. For example, Wuest
et al. found a relationship between chronic pain and severity of abuse, but only
for psychological abuse. Dutton et al. found that women who had experienced
more severe psychological and physical abuse were at greater risk of revictimization, but only if they had not also experienced severe sexual violence from
their partners. Other studies have found that particular forms of abuse, such as
harassment, put women at greater risk of revictimization, regardless of abuse
severity (e.g., Cole, Logan, & Shannon, 2008). These studies show that the
complex relationship between type and severity of abuse need to be taken into
account when measuring IPV and related outcomes.
Many health organizations (e.g., American Medical Association, 2008;
Joint Commission for the Accreditation of Healthcare Organizations [JCAHO],
2011) recommend screening for IPV in health care settings, although the evidence on whether screening does more good than harm is limited (Feder et al.,
2009) with a recent Canadian trial showing no beneficial health outcomes
from screening alone (MacMillan et al., 2009). The primary care setting has
potential to facilitate early intervention given the high level of contact women
have with family practice (Feder et al., 2011). The Women’s Evaluation of
Abuse and Violence Care (WEAVE) project is the first family practice based
trial testing the effect of screening plus intervention for IPV on women’s
health and wellbeing (Hegarty et al., 2010). The trial protocol (Hegarty et al.,
2010), intervention (Hegarty, O’Doherty, Gunn, Pierce, & Taft, 2008), and
screening stage (Hegarty, O’Doherty, Astbury, & Gunn, 2012) have been
described in detail elsewhere. In this article, we report an analysis of associations of type and severity of IPV with trial participants’ health and wellbeing,
safety planning, and use of support and clinical services. This fills a gap in
276
Journal of Interpersonal Violence 28(2)
the literature examining associations specifically of type and severity of IPV
with health outcomes.
Method
Between June 2008 and May 2010, women aged 16–50 years attending participating urban and rural general practitioners (GPs) in the previous 12 months
were sent a postal screening survey. Women were excluded if the GP anticipated
they may encounter difficulties in providing informed consent or understanding
the surveys. The survey contained a series of items about lifestyle issues
(Goodyear-Smith, Arroll, & Coupe, 2009), including an item about fear of
partner to screen for IPV (Hegarty, Bush, & Sheehan, 2005). Women who
indicated fear of a partner and supplied contact details were phoned by a
research assistant and invited to participate in the trial. Additional exclusions
were undertaken at this point if women had misinterpreted the fear item or
been fearful more than 12 months ago, had moved, or had poor English language skills. Women willing to be involved were sent a baseline survey, an
information leaflet and a resource services card to their nominated “safe”
address.
The baseline survey was adapted from surveys used in the DIAMOND
study (Gunn et al., 2008). The survey included socio-demographic items,
childhood abuse (Child Maltreatment History Self-Report; MacMillan et al.,
1997), quality of life (WHOQOL-Bref; Skevington, Lotfy, & O’Connell,
2004), physical and mental health status (Short-Form 12; SF-12; Ware,
Kosinski, & Keller, 1996), safety (Safety-Promoting Behaviour Checklist;
SPBC; McFarlane, Parker, Soeken, Silva, & Reel, 1998), anxiety and depression (Hospital Anxiety and Depression Scale; HADS; Zigmond & Snaith,
1983), posttraumatic stress disorder (Short Screening Scale for DSM-IV
PTSD; Breslau, Peterson, Kessler, & Schultz, 1999), days out-of-role; alcohol
use (Alcohol Use Disorders Identification Test; AUDIT; Saunders, Aasland,
Babor, de la Fuente, & Grant, 1993), smoking, and medication use.
IPV was measured using the Composite Abuse Scale (CAS; Hegarty et al.,
2005), which categorizes respondents according to four different types of abuse
experienced in the last 12 months: (i) Severe Combined Abuse, (ii) Physical
and Emotional Abuse, (iii) Physical Abuse only, and (iv) Emotional Abuse and/
or Harassment only. Women who did not meet the criteria for any of the abuse
types on the CAS (e.g., because the abuse was longer than 12 months ago) were
classified as fearful in the past 12 months but negative for IPV. The survey also
measured GPs’ inquiry about safety of women and their children, service utilization, and visits to health care providers.
Hegarty et al.
277
Once baseline surveys were returned, women were formally enrolled in
the trial and GPs (and women) were randomly allocated to intervention or
comparison groups. This article focuses on data emerging from the baseline
assessment only. The study was granted ethics approval by the Human
Research Ethics Committee of The University of Melbourne and is undertaken
in accordance with CONSORT guidelines (Schulz, Altman, & Moher, 2010).
Data were summarized using frequencies and percentages for categorical
data, and means and standard deviations for continuous data. Data met all
assumptions for statistical tests performed. Marginal logistic regression was
used to compare patient characteristics between the fearful women who
agreed to enrol in the trial and those who did not. Women’s experiences of
quality of life, health status, lifestyle, safety and help seeking behaviors and
health care utilization were compared across the four abuse categories
(namely, fearful but CAS-negative, Severe Combined Abuse, Physical and
Emotional Abuse, and Emotional Abuse, and/or Harassment only). Women in
the Physical Abuse only group were excluded from the analysis because there
were only five women. The association between each of the outcome variables and type of abuse was tested using marginal linear and logistic regression models for continuous and binary outcomes, respectively. The marginal
models were fitted using generalized estimating equations with information
sandwich estimates of the standard errors to allow for the correlation between
responses of the women that attend the same general practice.
Results were presented as mean score differences (95% confidence intervals, CI) for the continuous outcomes for each abuse category (Severe
Combined Abuse, Physical and Emotional Abuse, and Emotional Abuse and/
or Harassment only) compared to the group where women were fearful but
CAS negative, which was chosen as the reference category for the analysis.
Results were presented as odds ratios (OR) with 95% CIs for binary outcomes
for each of the three CAS positive groupings compared to the reference group.
Overall p-values that summarize the strength of the association between each
outcome and abuse type were also reported. Estimates were also adjusted for
age, marital status, income, and education based on a priori hypotheses of an
independent association with abuse derived from the literature (Hegarty,
Gunn, Chondros, & Small, 2004; Hegarty, Gunn, Chondros, & Taft, 2008).
Data were analyzed using STATA, version 11.0 (StataCorp, 2009).
Results
Figure 1 shows the flow of patient participants up to enrolment in the baseline phase of the trial. Across 55 practices initially recruited into the study,
278
Journal of Interpersonal Violence 28(2)
5,742/19,879 (28.9%) women returned a screening survey; 731/5,742
(12.7%) of them reported being afraid of a partner or ex-partner in the previous 12 months and 477/731 (65.3%) supplied contact details. Furthermore,
91 women were established as ineligible once they were contacted.
Ultimately, 272/386 (70.5%) women from 52 practices completed the baseline assessment and enrolled in the trial. Three GPs were excluded from the
analysis because none of their screened patients were eligible for participation. There was no difference between women who enrolled in the trial and
fearful women who did not enrol, by age, employment, relationship status,
or birthplace (see Table 1). However, trial enrolees were more often separated or divorced, had less school education, and were more likely to have
English as their first language compared to fearful non-enrolees. Trial enrolees also experienced more fear of partner than fearful non-enrolees (52.2%
versus 38.1% fearful some, most or all of the time).
Figure 2 shows the proportions of trial enrolees at baseline (N = 272) who
experienced each type of abuse in the last 12 months. Severe Combined
Abuse was the most common type of abuse experienced (33.0%), followed
closely by Physical and Emotional Abuse (26.2%), and Emotional Abuse
and/or Harassment only (26.6%); whereas only five participants (2.7%) had
experienced Physical Abuse only. More than half (64.0%) of trial participants
had experienced physical and/or sexual abuse as a child. Participants had fair
or poor health (31.6%) and poor social support (67.9%). They had high levels
of mental ill health, with one third of the women being on antidepressants
(33.2%) or painkillers (32.1%) and were frequent users of health care services. However, in the past 12 months, only 11.2% had seen a social worker,
14.7% had sought help from a domestic violence service, 20.4% had contact
with police, and 25.4% had contacted legal services. Around one quarter
(24.4%) had ever made a safety plan.
Table 2 shows that trial enrolees had lower scores on all dimensions of
quality of life (WHOQOL-Bref) and mental health (SF-12) compared to the
Australian population (Australian Bureau of Statistics, 1997; Hawthorne,
Herrman, & Murphy, 2006). Overall, women experiencing Severe Combined
Abuse reported lower quality of life on the physical, psychological, and environmental dimensions of the WHOQOL-Bref compared to all other abuse
categories. Although women experiencing some form of abuse (Severe
Combined Abuse, Physical and Emotional Abuse, or Emotional Abuse and/or
Harassment only) reported lower quality of life on the social dimension of the
WHOQOL-Bref and lower mean mental health scores on SF-12 compared to
the women who reported being fearful but CAS negative.
279
Hegarty et al.
Eligible patients sent screening survey (N=19879 at 55 practices)
Did not return survey (N=14137)
Returned survey (N=5742; 28.9%)
Screened positive for fear of partner
(N=731; 12.7%)
Did not screen positive for fear of partner
(N=5011; 87.3%)
Did not supply contact details (N=254)a
2 GPs excluded:
No eligible women identified
Supplied contact details
(N= 477 at 53 practices; 65.3%)
Declined: N=39a
Ineligible: N=91a
Could not be contacted: N=19a
Missed screening cut-off date: N=1a
Did not return baseline survey: N=55a
1 GP excluded:
No eligible women identified.
Returned baseline and enrolled in trial
(N=272 at 52 practices)
Figure 1. Flowchart of patient participation in the WEAVE trial before
randomization of practices
Note. aPatients who screened positive for fear of partner but did not enrol in the trial are
referred to as fearful non-enrolees throughout the article (N = 459).
Table 3 shows that women in the Severe Combined Abuse group were also
more likely than the other abuse groups to report PTSD, anxiety symptoms,
sedative, analgesic, and antidepressant use, and to experience days out-ofhome duties because of emotional issues. Those experiencing Severe
Combined Abuse or Physical and Emotional Abuse were also more likely
280
Journal of Interpersonal Violence 28(2)
Table 1. Socio-Demographic Characteristics of Trial Enrolees (N = 272) and NonEnrolees (N = 459)a
Age
Purchase answer to see full
attachment