Physical and Psychological Symptoms in Emotionally Abused and Non ...

Physical and Psychological Symptoms in
Emotionally Abused and Non-abused Women
John H. Porcerelli, PhD, Patricia A. West, RN, PhD, Juliann Binienda, MA, and
Rosemary Cogan, PhD
Background: This brief report compares emotionally abused and non-abused female family practice
patients on physical and emotional symptoms, alcohol use problems, and social support problems.
Methods: We conducted a secondary analysis of data from a cross-sectional, multicenter study of vic-
timization of family practice patients. Forty-seven adult women meeting criteria for emotional abuse
(within the past year) and no physical abuse were matched demographically with 47 non-abused
women. Each woman completed demographic and health history questionnaires, including questions
about physical and emotional abuse.
Results: Emotionally abused women reported a greater number of physical (P < .001) and psycho-
logical (P < .0001) symptoms than non-abused controls. Emotionally abused women reported a signifi-
cantly greater number of social support problems than non-abused women (P < .04).
Conclusions: This study supports a growing literature that demonstrates an association between
emotional abuse and physical and emotional symptoms in women who are currently suffering emotional
abuse at the hands of their partner or ex-partner. It is recommended that physicians inquire about
emotional abuse in female patients with multiple psychosocial and physical symptoms. (J Am Board Fam
Med 2006;19:201– 4.)
The negative health consequences of physical abuse
there is a subset of women who report only emo-
of women by their intimate partners are well doc-
tional abuse. Longitudinal studies suggest that
umented throughout the health science literature.
emotional abuse almost always precedes physical
Less is known about the consequences of emotional
abuse.2 In studies of women who report both phys-
abuse to women. Emotional abuse (also referred to
ical and emotional abuse, negative health outcomes
as psychological abuse) can include any one of the
are as strongly associated with emotional abuse as
following alone or in combination: threats of phys-
they are with physical abuse.3,4
ical harm, physical and social isolation, extreme
The present study compared female family med-
jealousy and controlling behavior, degradation, in-
icine patients reporting emotional abuse (and no
timidation and other forms of chronic verbal ha-
physical abuse by a partner, ex-partner, or non-
rassment, withdrawal, destroying trust, and placing
partner) with a group of non-abused women
a partner in a dangerous situation.1 Although most
matched for age, race, income, employment, and
women who suffer physical abuse at the hands of
education. We compared emotionally abused and
their partner also report being emotionally abused,
non-abused women on self-reported physical
symptoms and psychosocial variables. We hypoth-
esized that emotionally abused women will have
Submitted 3 May 2005; revised 26 July 2005; accepted 29
more physical and psychological symptoms, alcohol
July 2005.
use problems, and social support problems than
From the Department of Family Medicine, Wayne State
non-abused women.
University School of Medicine, Detroit, MI (JHP, JB); Fam-
ily Practice Residency, St. John Hospital and Medical Cen-
ter, St. Clair Shores, MI (PPW); and Department of Psy-
chology, Texas Tech University, Lubbock, TX (RC).
Methods
Funding: This work was supported by grants from the
Oakland Health Education Program (OHEP) Center for
In this cross-sectional group comparison study de-
Medical Education (Southfield, MI) and St. John Hospital
sign, participants were obtained from a multicenter
and Medical Center Research Fund (St. Clair Shores, MI).
prevalence study of violent victimization of male
Conflict of interest: none declared.
Corresponding author: John H. Porcerelli, PhD, Depart-
and female family practice patients.5 In the original
ment of Family Medicine, Wayne State University School of
study, 713 women and 350 men from 4 family
Medicine, 15400 W. McNichols—2nd Floor, Detroit, MI
48235 (E-mail: jporcer@med.wayne.edu).
practice clinics (1 urban and 2 suburban residency
http://www.jabfm.org
Emotional Abuse 201

---

training sites and 1 suburban faculty practice) par-
tercourse, vaginal bleeding after intercourse, pelvic
ticipated and were screened for physical and emo-
pain), skin (skin problems or changes in your skin),
tional abuse. Screening was conducted with con-
aching muscles or joints. Women reporting both
secutive female patients on 3 half-days per week in
physical and emotional abuse within the past year
each of the clinic waiting rooms for a duration of
were excluded from this study. Psychological symp-
2 months. Only data on health correlates of phys-
toms include 6 depression items (exhausted or fa-
ical victimization were reported in that study. Of
tigued most of the time, felt blue, lonely or de-
the 713 women, 47 adult women (ages 18 to 64)
pressed, more irritable than usual, frequent crying
reported being emotionally abused by their partner
spells, suicidal ideation) and 2 anxiety items (diffi-
“within the past year” (not physically abused) and
culty trying to calm down or relax and overly anx-
thus are the focus of this study. A comparison
ious or worrying a lot). Alcohol use problems in-
group included 47 women matched for age ( 5
clude the 4 CAGE7 questions and an additional
years), race (97% exact match), education (94%
item on quantity of daily use. Four social support
exact match), employment (89% exact match), and
items include (time well-balanced between work,
income (83% exact match). Non-exact matches
family, and play; relationship with friends; relation-
were within one level of the standard demographic
ship with partner; someone to discuss personal
categories. Each emotionally abused woman was
problems with).
matched with one non-abused woman from our
Inclusion criteria for the emotional abuse group
original sample. When more than one non-abused
included women who reported being emotionally
woman was eligible for matching to a woman in the
abused by their partner or ex-partner within the
emotionally abused group, one was randomly cho-
previous year and did not indicate physical abuse
sen using a random number generator.
within the previous year by partner, ex-partner, or
Participants responded to a demographic ques-
non-partner. Women in the control group did not
tionnaire, the Brief Conflict Tactics Scale6 a single
report either physical or emotional abuse within
question with adequate validity for screening phys-
the past year.
ical abuse in emergency department settings (Have
Because the groups in our study were matched
you been hit, kicked, punched or otherwise hurt by
on several demographic variables and thus were not
someone in the past year? If so, by whom?), a face-valid
considered as independent groups, paired-sample t
question about emotional abuse developed by the
tests were conducted comparing the emotionally
investigators (Have you felt controlled, threatened, or
abused group with the matched control group on
afraid of someone within the past year? If so, by whom?),
each of the main dependent variables (physical
and a checklist of perpetrators (family member,
symptoms, psychological symptoms, alcohol use
friend, partner/ex-partner, stranger). If a patient
problems, and social support problems).
checked the “yes” box for either abuse question,
they were also requested to place a check mark next
to the perpetrator(s).
Results
In addition, a modified 88-item version of the
Because the groups in this study were matched
Milcom Health History Update and Physical Ex-
according to age, race, education, and income,
amination form developed by Hollister, Inc. in co-
overall scores are reported for the total sample:
operation with the Society of Teachers of Family
mean age
35.72 (SD
9.83; range 18 to 59);
Medicine was administered. The Milcom is made
Caucasians
52%, African Americans
42%,
up of standard physical and emotional health items
other
6%; 72% of patients had a family income
answered in a “yes-no” format. For this study, 17
less than $51,000, and 87% had at least a high
physical symptoms included head, ears, eyes, nose,
school education.
and throat (HEENT) items (headaches, dizziness,
The results of the analyses comparing emotion-
seizures, troubles with your ears, dental or other
ally abused and non-abused women on physical and
mouth problems, and nose bleeds), respiratory/
psychological symptoms, alcohol use problems, and
cardiovascular items (palpitations and chest pain),
social support problems are reported in Table 1.
gastrointestinal (abdominal discomfort and pain,
The t values for physical symptoms (P
.001),
nausea or vomiting, difficulty swallowing), genito-
psychological symptoms (P
.0001), and social
urinary (menstrual changes, discomfort during in-
support (P
.043) were statistically significant,
202 JABFM March–April 2006 Vol. 19 No. 2
http://www.jabfm.org

---

Table 1. Physical and psychological symptoms, alcohol use problems and social support problems in emotionally
abused and non-abused women.
Emotional Abuse
Non-abused Matched
Group N
47
Controls N
47
t Values*
Effect Size
Symptoms
Mean
SD
Mean
SD
t
P
Cohen’s d
Physical symptom total
5.47
3.62
3.13
2.70
3.46
.001
0.73
Psychological symptom total
4.66
2.24
2.57
2.17
4.34
.0001
0.94
Alcohol use problem total (CAGE)
0.42
0.90
0.34
0.96
1.94
.45
0.09
Social support problem total
1.49
0.83
1.11
0.87
2.08
.043
0.43
* Degrees of freedom for all 4 paired-sample t tests
46.
thus supporting 3 of 4 hypotheses. Because statis-
and psychological symptoms.2–4 These studies
tical significance is affected by sample size, effect
showed symptoms in multiple systems including
sizes were also used in this study to assess the
neurological, cardiovascular, abdominal, and geni-
strength of the differences between groups. Ac-
tourinary as well as in psychological (mostly de-
cording to Cohen’s criteria, an effect size of 0.20 is
pressive) symptoms. Women in the emotional
considered to be a small effect, 0.50 is considered a
abuse group also reported a greater number of
medium effect, and 0.80 is considered a large ef-
negative social support items, ie, changes in their
fect.8 In this study, differences in psychological
relationship with their partner within the past year.
symptoms between the matched groups evidenced
We hypothesized that emotionally abused
a large effect (0.95), whereas the differences in
women have higher alcohol use problem scores
physical symptoms approached a large effect (0.74),
than non-abused women (ie, emotionally abused
and social support problems (0.43) evidenced a
women would turn to alcohol as a coping mecha-
moderate effect. Thus despite the small sample
nism). This was not true in our study sample. In our
sizes of the groups, substantial differences, espe-
original study,5 physically victimized women evi-
cially in physical and emotional symptoms, were
denced less alcohol use problems than physically
obtained. Group differences approaching large ef-
victimized men. However, women who were phys-
fect sizes are likely to have clinical significance.
ically victimized by more than one type of perpe-
Exploratory analyses were conducted to deter-
trator (eg, partner and stranger) evidenced more
mine which specific physical symptom differed be-
alcohol use problems than women who were phys-
tween the emotionally abused and non-abused
ically victimized by a single perpetrator or non-
groups. The t tests for each of the items revealed 8
victimized.
of 18 physical symptoms were significantly greater
The need for identifying physical abuse of
in the emotionally abused group: HEENT items
women in primary care is well established. How-
related to dizziness and seizures, cardiac items re-
ever, the findings from this and other recent stud-
lated to palpitations and chest tightness, gastroin-
ies,2–4 indicate that physicians should inquire about
testinal items related to abdominal discomfort and
emotional abuse in women who present with mul-
difficulty swallowing, and genitourinary items re-
tiple physical and psychological symptoms. Except
lated to change in menstrual periods. Results of the
for
the
HITS
(hurt,
insulted,
threatened,
analyses of specific psychological symptoms re-
screamed),9 instruments used for identifying part-
vealed differences on all items except for suicidal
ner abuse in primary care settings include physical
ideation. Only one of the social support problem
victimization items and rarely include an item
items differed between the emotional and non-
about emotional abuse. For example, the Patient
abuse groups: “Is your relationship with your
Health Questionnaire,10 a valid self-report psychi-
spouse/partner as good as it was last year?”
atric diagnostic instrument designed for primary
care settings, includes one domestic violence item
Discussion
having to do with physical abuse only. It is recom-
This study supports a growing literature on the
mended that both physical and emotional abuse
relationship between emotional abuse and physical
items be included in standard assessment scales.
http://www.jabfm.org
Emotional Abuse 203

---

Both self-report and patient-centered interviewing
ing critical attention in domestic violence. Violence
can aid primary care physicians in providing com-
Vict 1999;14:3–23.
prehensive preventive health care to their female
3. Coker AL, Smith PH, Bethea L, King MR, McNe-
own RE. Physical health consequences of physical
patients as it relates to interpersonal violence.
and psychological intimate partner violence. Arch
Limitations of the study include a cross-sec-
Fam Med 2000;9:451–7.
tional design which limits our ability to indicate a
4. Coker AL, Davis KE, Arias I, et al. Physical and
causal link between emotional abuse and physical
mental health effects of intimate partner violence for
and psychological symptoms. Longitudinal studies
men and women. Am J Prev Med 2002;23:260 – 8.
comparing pre- and post-abuse physical and psy-
5. Porcerelli JH, Cogan R, West PP, et al. Violent
chological status are needed with primary care pa-
victimization of women and men: physical and psy-
chiatric symptoms. J Am Board Fam Pract 2003;16:
tients. Additional limitations of the study include a
32–9.
small sample size, lack of statistical adjustment of
6. Feldhaus KM, Kozio-McLain J, Amsbury HL, Nor-
common confounding variables (eg, degree of so-
ton IM, Lowensein SR, Abott JT. Accuracy of three
matization, history of childhood abuse/neglect,
brief screening questions for detecting partner vio-
etc), a reliance on self-reported physical symptoms,
lence in the emergency department. JAMA 1997;
and the use of a single (global) emotional abuse
277:1357– 61.
screening item for the determination of emotional
7. Ewing JA. Detecting alcoholism: the CAGE ques-
tionnaire. JAMA 1984;252:1905–7.
abuse. However, the success of this single item in
8. Cohen J. Statistical power analysis for the behavioral
the present study warrants further study of its con-
sciences. 2nd Ed. Hillsdale (NJ): Lawrence Erlbaum
vergent and predictive validity with other known
Associates, Inc; 1988.
scales of emotional abuse.
9. Sherin KM, Sinacore JM, Li X, Zitter RE, Shakil A.
HITS: a short domestic violence screening tool for
use in a family practice setting. Fam Med 1998;30:
References
508 –12.
1. Eyler AE, Cohen M, Kershaw MO. Domestic vio-
10. Spitzer RL, Kroenke K, Williams JB. Validation and
lence and abuse. In: Knesper DJ, Riba MB, Schwenk
utility of a self-report version of PRME-MD: The
TL, editors. Primary care psychiatry. Philadelphia
PHQ primary care study. Primary care evaluation of
(PA): WB Saunders; 1997. p. 387– 68.
mental disorders. Patient Health Questionnaire.
2. O’Leary DK. Psychological abuse: a variable deserv-
JAMA 1999;282:1737– 44.
204 JABFM March–April 2006 Vol. 19 No. 2
http://www.jabfm.org

---