Lövestad et al.

BMC Public Health (2017) 17:335

DOI 10.1186/s12889-017-4222-y

RESEARCH ARTICLE Open Access

Prevalence of intimate partner violence and

its association with symptoms of
depression; a cross-sectional study based
on a female population sample in Sweden
Solveig Lövestad1* , Jesper Löve1, Marjan Vaez2 and Gunilla Krantz1

Abstract
Background: Intimate Partner Violence (IPV) is the most common type of violence targeting women. IPV includes
acts of physical aggression, sexual coercion, psychological abuse and controlling behaviors and these forms of
violence often coexist in the same relationship. Living with IPV is associated with serious mental health outcomes
such as depression and depressive symptoms. Few population based studies from Sweden have investigated the
relationship between different forms of IPV and women’s depressive symptoms and even fewer used controlling
behavior as an independent variable in such studies. The aim of this study was therefore to assess the prevalence
of exposure to IPV in terms of controlling behavior, sexual, and physical violence and their association with self-
reported symptoms of depression in a female population based sample.
Methods: The cross-sectional, population based sample contained 573 women aged 18–65 years randomly
selected in Sweden. Five self-reported symptoms that define depression in the Diagnostic and Statistical Manual of
Mental Disorders were assessed. Physical and sexual violence were inquired about using the World Health
Organization’s (WHO) Violence Against Women Instrument (VAWI), while controlling behavior was assessed with the
Controlling Behavior Scale (CBS). Associations between different forms of IPV and symptoms of depression were
estimated by crude and adjusted odds ratio (OR) with 95% confidence intervals (CI).
Results: Bivariable associations revealed that women exposed to controlling behavior, had higher OR of depressive
symptoms compared to unexposed women (OR 2.43; 95% CI 1.63–3.63). Women exposed to physical and sexual
violence had also a higher OR of depressive symptoms (OR 3.78; 95% CI 1.99–7.17 and OR 5.10; 95% CI 1.74–14.91
respectively). After adjusting for socio-demographic and psychosocial covariates, all three forms of IPV showed
statistically significant associations with self-reported symptoms of depression.
Conclusions: A strength with this study is the analysis of controlling behavior and its association with self-reported
symptoms of depression in a female population based sample. Exposure to controlling behavior, physical and
sexual violence by an intimate partner were clearly associated with women’s self-reported symptoms of depression.
Keywords: Intimate partner violence, Depression, Population-based, Women, Sweden

* Correspondence: [email protected]
1
Department of Community Medicine and Public Health, Sahlgrenska
Academy at University of Gothenburg, Box 453, 405 30 Göteborg, Sweden
Full list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Lövestad et al. BMC Public Health (2017) 17:335 Page 2 of 11

Background economic burden in terms of disability, sick-leave and

Intimate Partner Violence (IPV) perpetrated by a current health care costs [25].
or former partner is the most common type of violence Many studies have reported on IPV as an overarching
targeting women [1] and continues to be a gross violation construct, i.e. not separating the various forms of IPV
of women’s human rights as well as a major public health when analyzing its impact on depressive symptoms
problem globally [2]. The World Health Organization [21, 26–28]. Despite the fact that psychological abuse is
(WHO) defines IPV as: ‘any behavior within an intimate more prevalent than other forms of IPV, most studies have
relationship that causes physical, sexual or psychological focused solely on physical and sexual IPV as exposure vari-
harm, including acts of physical aggression, sexual coer- ables [22]. Johnson and colleagues [29] for example, found
cion, psychological abuse and controlling behaviors’ [3]. in their longitudinal study that exposure to physical IPV
The term ‘controlling behavior’ includes acts that restrict was associated with depressive symptoms. Likewise, sexual
a woman’s mobility or her access to relatives and friends IPV has been identified as an important and independent
while ‘psychological abuse’ refers to threats, insults, and risk factor for later depression in women [22, 30]. Further-
acts that belittle or humiliate the partner [4]. However, in more, studies that included psychological abuse as a separ-
most studies controlling behavior is viewed as a form of ate and independent variable have with few exceptions,
psychological violence [5] and this is therefore inquired demonstrated a consistent relationship with depression
about and analyzed as a unitary construct [6–8]. The im- [6, 15, 31]. Additional factors known to be associated
portance of distinguishing controlling behavior from other with depressive symptoms are having poor social
acts of psychological violence has repeatedly been empha- support [21], younger age [32], being single [21], being
sized by Johnson [9], as a form of IPV that has devastating unemployed [33] and having witnessed inter-parental
health and social consequences both in itself and in com- IPV [34].
bination with physical and/or sexual violence. There is limited research on the association between
A recently published report from the 28 Member different forms of IPV and women’s depressive symp-
States of the European Union (EU) showed that over toms in population based samples in Sweden. Earlier
one in five women in the EU had experienced physical studies on this matter were mainly conducted on clinical
and sexual violence from either a current or former in- samples [35–40] or on specific target groups, e.g., Thai
timate partner and 35% had experienced controlling be- women, or women from shelters [41, 42], i.e. studies not
havior [5]. In one of our earlier studies performed in representative of the general female population. Many of
Sweden, 8% of the women reported exposure to physical the earlier studies addressed the association between
IPV during the past year while 3% reported exposure to violence victimization and mental health without asking
sexual IPV [10]. Another study performed in Sweden the respondent to specify whether the perpetrator had
showed that almost 2% of the female respondents had been an acquaintance, a stranger, a family member or an
experienced systematic and repeated acts of controlling intimate partner, thus limiting the possibility to distin-
behavior during past year prior to the survey [11]. guish between IPV and other forms of violence [43–45].
Previous research has shown that various forms of IPV Since IPV is the most common type of violence against
generally coexist in the same relationship [12–14]. How- women [1], it is important to provide information about
ever, some studies suggest that psychological abuse, includ- its forms and consequences in a general population-
ing controlling behavior, is far more frequent than other based sample of women in Sweden as preventive mea-
forms of IPV [14, 15] and that most women exposed to sures and treatment differ considerably depending on
physical IPV, also are exposed to some form of psycho- who is the perpetrator. To the best of our knowledge,
logical abuse [16, 17]. A study performed in Sweden for in- few if any studies performed in Sweden include control-
stance, showed that four out of ten women who reported ling behavior as an independent variable in studies on
exposure to jealousy from their partner, also reported ex- exposure to IPV and its association with symptoms of
posure to physical and sexual violence [18]. depression among women.
In addition to physical injury, it has repeatedly been The aim of this study was therefore to assess the
demonstrated that IPV is associated with mental health prevalence of exposure to IPV in terms of controlling
problems, including depression and depressive symp- behavior, sexual and physical violence and its association
toms [19–22]. Depression is the most frequent mental with self-reported symptoms of depression in a female
health problem among women and is twice as common population based sample in Sweden.
in women as in men [23]. Symptoms and severity of
depression varies largely and may include self-reported Methods
measures as well as diagnoses based on the Diagnostic Design and sample
and Statistical Manual (DSM) [24]. Apart from individ- The present cross-sectional study was based on survey
ual suffering, depression leads to high societal and data extracted from a larger programme on exposure and
Lövestad et al. BMC Public Health (2017) 17:335 Page 3 of 11

perpetration of IPV among men and women in Sweden study, was self- reported experience of exposure to con-
[46]. Between January and March 2009, a postal survey trolling behavior, physical and sexual violence during
administered through Statistics Sweden was sent out to a past 12 months and perpetrated by a current or former
random national sample of 1006 women and 1009 men, partner including spouses, common-law partners or
aged 18–65 years residing in Sweden. A short letter with boy/ girlfriends within opposite and same-sex relation-
information about the study and possibility to deny fur- ships. The WHO Violence Against Women Instrument
ther participation was sent out to all selected individuals (VAWI) was used to assess physical (6 items) and sexual
one week before the questionnaire was sent out. This was (3 items) violence [48] perpetrated by an intimate partner
done as a security measure, to avoid aggravating possible (Table 1). VAWI demonstrated good construct validity
violence exposure. The respondents were guaranteed full and internal reliability in one of our earlier studies per-
anonymity. For further information and/or assistance in formed on this sample [49]. Questions on exposure to
case of being exposed to IPV, respondents were given con- controlling behavior were measured through the subscale
tact details to a general practitioner, a psychologist and a ‘isolating control’ (5 items) (Table 1) from the Controlling
contact person at Statistics Sweden. The study included Behavior Scale (CBS) developed by Graham-Kevan and
two reminders in order to minimize the drop-out rate. Archer [50]. The original 5-point response format ranging
A total of 624 women (62.0%) and 458 men (45.5%) from 0 (never) to 4 (always) [50] was modified into fre-
returned the questionnaire. Respondents with missing quency questions. For each question on physical, sexual
values (n = 110) on all the violence items were excluded, violence and controlling behavior, respondents were asked
leaving a final sample of 972 men and women. The how often they had experienced a specific act during the
current study is based on 573 women who formed the past 12 months. The response options were ‘0 times’, ‘1
final sample. Earlier drop-out analysis on this sample time’, ‘2 times’, ‘3–5 times’ or ‘> 5 times’. For each item, ex-
found that women born outside Sweden, of younger age, posure was considered as present if the respondent had
unmarried and with a low annual income were over- experienced violent behavior ‘1 to >5 times’ during the
represented in the group of non-responders [46]. past year. Due to few cases in each of the frequency cat-
egories the cumulative number was used in the further
Variables analyses. For further bivariable and multivariable analysis,
Outcome variables the three forms of violence were analyzed separately.
Self-reported symptoms of depression were assessed using Items included in each of the three subscales (‘isolating
five indicators of depression as defined in the Diagnostic control’, physical- and sexual violence), were summarized
and Statistical Manual of Mental Disorders, Fourth Edi- and dichotomized into a binary variable where exposure
tion (DSM- IV) [47]. Indicators of depressive symptoms to violent behavior was defined as having reported expos-
according to the DSM-IV include feelings of sadness and ure to at least one of the items in each subscale as op-
discouragement, initial insomnia, decreased energy and posed to experiencing no such violence.
tiredness as well as impaired ability to concentrate [47].
Symptoms of depression further encompasses suicidal Covariates
ideation and attempts [47]. The respondents were asked if For descriptive analysis Age was divided into five age in-
they had experienced any of the following five symptoms tervals; 18–25, 26–35, 36–45, 46–55 and 56–65 years
during the past 12 months: ‘Tiredness/ fatigue’, ‘difficulty and for further analysis categorized into 18–25 years and
falling asleep’, ‘trouble concentrating’, ‘feeling down/ low’ 26–65 years with the latter as reference category. Civil
and ‘suicidal thoughts’. A four point scale (‘Almost every status was categorized into three groups: (1) single,
day’, ‘Once a week’, ‘Once a month’ and ‘Almost never or widowed, divorced; (2) boyfriend, girlfriend; (3) married, co-
never’) was used to indicate the frequency of the various habitant, registered partnership, and later dichotomized by
depressive related symptoms experienced in the past year. merging the two former categories. Educational level was
The response options ‘almost every day’ and ‘once a week’ categorized into three groups (university, high school and
were merged and considered as exposed. For the bivari- elementary school). The categories ‘university/high school’
able and multivariable analysis, the five items were sum- were combined and used as the reference category. Employ-
marized and dichotomized into the exposure category, ment status was grouped into seven categories: (1)
defined as two or more out of five symptoms and the ref- employed; (2) student; (3) early retirement pension/retired;
erence category of having one or no such experience of (4) sick leave; (5) parental leave or leave of absence; (6) un-
depressive symptoms. employed; (7) home-worker/taking care of the household.
The categories were dichotomized, using ‘employment’ and
Exposure variables ‘parental leave or leave of absence’ as the reference group.
The different forms of IPV were analyzed as primary ex- The measure ‘Access to social support’ has been used in
posure variables. Being exposed to IPV, as defined in this the Swedish Level of living surveys (LNU) [51] and was
Lövestad et al. BMC Public Health (2017) 17:335 Page 4 of 11

Table 1 Exposure to intimate partner violence presented as past year frequency (n) and percentage (%) of the total population.
N = 573 women
Intimate partner violence Exposure during past 12 months
% n
Controlling behavior
Tried to restrict time spent with my family and friends 4.9 28
Wanted to know where I went and who I spoke to when not together 17.6 101
Tried to restrict my activities outside the relationship 6.4 37
Felt suspicious and jealous of me 12.6 72
Tried to control my activities 6.6 38
Exposed to ≥ 1 item of controlling behavior 25.0 143
Physical violence
Pushed or shoved me 6.6 38
Thrown something that could have hurt me 1.6 9
Hit me with the fist or with some other object 1.2 7
Kicked and dragged me and beaten me up 0.5 3
Choked or burnt me on purpose 0.5 3
Hurt me with a knife, a gun or some other weapon 0.2 1
Exposed to ≥ 1 item of physical violence scale 7.5 43
Sexual violence
Demanded to have sex with me even though I did not want to 2.4 14
Forced me to have sex against my will by using physical strength 0.3 2
Forced me to perform sexual acts that I experienced as degrading and/ or humiliating 0.3 2
Exposed to ≥ 1 item of sexual violence scale 2.8 16

constructed out of the question: ‘One sometimes needs Statistical analyses

help and support from someone. Do you have any relative Analyses were computed using the statistical program
or friend who helps out…’ followed by four subsequent SPSS, version 17 and 20. Descriptive statistics using
questions: (1) if you become ill?; (2) if you want company?; prevalence (%) and frequency (n) were used for preva-
(3) if you need to talk to someone about personal prob- lence rates. The overlap between physical, sexual vio-
lems?; (4) if you need a loan of 15,000 Swedish crowns (at lence and controlling behavior during past 12 months
that time approx. 2200 USD)? Answering ‘yes’ to all four was illustrated by a Venn-diagram. Bivariable and multi-
questions was categorized as ‘access to good social sup- variable analyses were performed producing crude and
port’ whereas answering ‘no’ or ‘unsure’ to any of the four adjusted odds ratios (OR) with 95% Confidence Intervals
items was considered as ‘poor social support‘. (CI) in order to analyze associations between different
Grown up and witnessed IPV as a child was con- forms of IPV, covariates and symptoms of depression.
structed out of two introductory questions followed by Exposure variables that showed statistically significant as-
more detailed questions. The first question; ‘Have you sociations with symptoms of depression in the crude ana-
grown up in a home where there was physical, psycho- lyses were entered one-by-one in a stepwise manner into
logical or sexual violence between your parents or the the hierarchical logistic regression analysis. Models were
adults you lived with?’ was followed by the response op- created with the composite measurement of symptoms of
tions ‘no’; ‘yes’ and ‘unsure’. If the answer was ‘yes’, the re- depression and each form of IPV. Despite non significance
spondent was asked to indicate which type of violence in the bivariable analyses, age was included in the multivar-
(physical, psychological and/or sexual violence). The sec- iable analyses because of its known association with both
ond question asked whether the respondent had witnessed depression [32] and IPV [5]. The first model was adjusted
(heard or seen) the violence or not. The response options for age, civil status and employment status and the second
were ‘no’; ‘yes’ or ‘unsure’. A binary variable was con- model was adjusted for age, civil status, employment status
structed in which women responding ‘yes’ to both intro- and access to social support. The third model was adjusted
ductory questions were considered to have grown up and for age, civil status, employment status, access to social sup-
witnessed violence and were thus defined as ‘exposed’. port and grown up with and witnessed IPV as a child.
Lövestad et al. BMC Public Health (2017) 17:335 Page 5 of 11

Internal reliability Table 2 Socio - demographic and psychosocial characteristics.

To assess the internal consistency of the items measuring N = 573 women
self- reported depressive symptoms, Cronbach’s α was Characteristics Percent% Number
computed to 0.76. Age groups
18–25 10.5 60
Ethical considerations 26–35 23.6 135
The current study was conducted in accordance with the 36–45 22.5 129
WHO’s ethical and safety recommendations for research
46–55 20.9 120
on IPV [52]. Approval was provided from the Regional
Ethics Review Board in Gothenburg (Dnr: 527–08). 56–65 22.5 129
Civil status

Results Single/ widowed/ divorced 14.8 85

Descriptive statistics Boyfriend/ girlfriend 11.2 64
The sample consisted of 573 women aged 18 to Married/ cohabitant/ registered partnership 73.3 420
65 years, with an average age of 42.7 years (Standard Duration of present relationship
Deviation =13.01). The majority of women (73.3%) were
> 10 years 50.3 288
married, cohabiting or in a registered partnership (Table
4-10 years 20.1 115
2) and seven (1.2%) women reported having a same-sex
relationship. Of the sample, 47.1% of the women had a ≤ 3 years 14.3 82
university degree and further 69.7% were employed. Of Country of birth
the respondents, 7.7% had grown up with- and witnessed Sweden 90.6 519
IPV as a child. Other Nordic country 2.6 15
Other European country 3.1 18
Prevalence, frequency and co-occurrence of controlling Country outside of Europe 3.7 21
behavior, physical and sexual violence
Educational level
As shown in Table 1, the most common form of violence
was controlling behavior (25.0%) during the past 12 months, University 47.1 270
followed by physical violence (7.5%) and sexual violence High school (10–12 years) 36.8 211
(2.8%). The most prevalent acts of controlling behavior Elementary school (≤ 9 years) 15.9 91
were ‘Wanted to know where I went and who I spoke to…’ Employment status
and ‘My partner felt suspicious and jealous of me’ with Employed 69.7 396
prevalence rates of 17.6% and 12.6% respectively. In total
Student 6.2 35
28% (n = 159) of the women were exposed to at least one
type of violence during the past 12 months (Fig. 1). Among Early retirement pension/Retired 8.2 47
the exposed women, 16.4% were exposed to controlling be- Sick leave (more than 3 months) 1.4 8
havior and physical violence while 4.4% of the women were Parental leave or leave of absence 6.2 35
subjected to both controlling acts and sexual violence. Ex- Unemployed 4.0 23
perience of all three forms of violence accounted for 2.5% Home-worker, taking care of the 4.2 24
(Fig. 1) of the respondents. household & other
Grown up with- and witnessed IPV
Prevalence and frequency of self-reported symptoms of No/ Unsure 91.3 523
depression Yes 7.7 44
Of the total sample, 31.6% of the women reported ex-
Access to social support
perience of at least two out of five symptoms of depres-
sion almost every day or once a week during the past Good 62.8 360
12 months (Table 3). Among the women 45.7% reported Poor 34.4 197
experiencing noticeable fatigue and tiredness every day or
once a week while 29.7% reported difficulties falling asleep Crude associations between IPV and self-reported symptoms
and further 18.3% had experienced difficulties in concen- of depression
trating during the past 12 months. Among the respon- Among women exposed to controlling behavior, 38.5%
dents, 0.7% (n = 4) had experienced suicidal thoughts reported at least two out of five symptoms of depression
almost every day or once a week throughout the year pre- almost every day or once a week (Table 4) and the odds
vious to the study. for depressive symptoms was 2.43 with 95% CI of 1.63–
Lövestad et al. BMC Public Health (2017) 17:335 Page 6 of 11

respectively) compared to women with no experience of

such violence. Likewise, being single/divorced/widowed,
being unemployed, having poor social support and having
66.7 (106) grown up with- and witnessed IPV as a child, all demon-
strated significant crude associations with self-reported
Control behavior
symptoms of depression (Table 4).

Adjusted associations between IPV and self-reported

symptoms of depression
After adjusting for age, civil status, employment status, ac-
16.4 (26)
4.4 (7) cess to social support and grown up with-and witnessed
IPV as a child, all three forms of IPV still showed statis-
2.5 (4) tical significance with self-reported symptoms of depres-
Physical violence sion (Model 3, Table 5). Women exposed to controlling
Sexual violence behavior during past 12 months were more likely to report
1.2 (2)
6.9 (11) symptoms of depression (OR 2.43; 1.56–3.79) compared
1.9 (3) to unexposed women. Likewise, women exposed to phys-
ical and sexual violence, had higher odds to report such
symptoms (OR 3.06; 1.50–6.24 and OR 4.67; 1.35–16.18
Fig. 1 Overlap of exposure to physical and sexual violence and respectively) compared to the reference categories.
controlling behavior (N = 159). Presented as past year frequency (n) It was also of interest to note that of the covariates an-
and percentage of women exposed to any type of violence alyzed in the multivariable analyses, all variables (single/
divorced, unemployed, poor access to social support,
3.63 as compared to the non-exposed. Women exposed grown up with-and witnessed IPV) except low age were
to controlling behavior alone (without physical or sexual associated with self-reported symptoms of depressive
violence), were more likely to report depressive symptoms disorders during past 12 months (Table 5).
(n = 43), compared to women without such experience
(OR 1.72; 1.10–2.67) (not shown in table). Further, women Discussion
exposed to physical violence and those exposed to sexual This is one of few population-based studies performed
violence had higher odds of reporting depressive symp- in Sweden that investigated the association between dif-
toms (OR 3.78; 1.99–7.17 and OR 5.10; 1.74–14.91 ferent forms of IPV, i.e. controlling behavior, physical
Table 3 Self-reported symptoms of depression presented as past year frequency (n) and percentage (%) of the total population.
N = 573 Women
Self-reported symptoms of Almost every day Once a week Once a month Almost never/ Never Exposed almost every
depression past 12 months day/once a week
% (n) % (n) % (n) % (n) % (n)
Fatigue/ tiredness 15.5 (89) 30.2 (173) 33.3 (190) 17.1 (98) 45.7 (262)
Difficulty falling asleep 8.7 (50) 20.9 (120) 26.7 (153) 40.5 (232) 29.7 (170)
Trouble concentrating 6.1 (35) 12.2 (70) 25.0 (143) 50.1 (287) 18.3 (105)
Feeling down/ low 6.3 (36) 11.3 (65) 31.2 (179) 44.9 (257) 17.6 (101)
Suicidal thoughts 0.5 (3) 0.2 (1) 3.5 (20) 89.0 (510) 0.7 (4)
Exposed to number of depression related symptoms (almost every day/ once a week) past 12 months
Number of depression related Exposed almost every
symptoms day/once a week
% (n)
1 symptom 23.7 (136)
2 symptoms 15.4 (88)
3 symptoms 7.9 (45)
4 symptoms 7.9 (45)
5 symptoms 0.5 (3)
At least 2 out of 5 symptoms 31.6 (181)
Lövestad et al. BMC Public Health (2017) 17:335 Page 7 of 11

Table 4 Bivariable associations between exposure to partner violence, covariates and self-reported symptoms of depression. Presented
as past year prevalence (n), percentage (%) and crude odds ratio (OR) and 95% confidence intervals (95% CI). N = 573 women
Explanatory variables and potential confounders Self-reported symptoms of depression past 12 months
Experienced ≥ 2 out of 5 symptoms Crude OR (95% CI)
% (n)
Controlling behavior past 12 months
Unexposed 61.5 (104) 1
Exposed (≥1 of the items) 38.5 (65) 2.43 (1.63–3.63)
Physical violence past 12 months
Unexposed 84.6 (143) 1
Exposed (≥1 of the items) 15.4 (26) 3.78 (1.99–7.17)
Sexual violence past 12 months
Unexposed 93.5 (158) 1
Exposed (≥1 of the items) 6.5 (11) 5.10 (1.74–14.91)
Age groups
26–65 86.7 (157) 1
18–25 13.3 (24) 1.49 (0.86–2.58)
Civil status
Married, cohabitant, registered partnership, boy- or girlfriend 77.2 (139) 1
Single, widowed, divorced 22.8 (41) 1.46 (1.53–3.96)
Educational level
University/ High school 83.9 (151) 1
Elementary school 16.1 (29) 1.00 (0.62–1.62)
Employment status
Employed, parental leave, leave of absence 64.1 (116) 1
Student, unemployed, sick-leave, early retirement /retired, home-worker, other 35.9 (65) 2.45 (1.64–3.64)
Social support
Good 50.3 (88) 1
Poor 49.7 (87) 2.45 (1.70–3.56)
Grown up with- and witnessed IPV
No/ Unsure 85.4 (152) 1
Witnessed violence 14.6 (26) 3.92 (2.05–7.52)

and sexual violence, and self- reported symptoms of de- biomedical findings suggest that sustained psychological
pression during past 12 months. stress due to social threat or rejection may up-regulate pro-
inflammatory cytokine activity which can alter the activity
Exposure to IPV and its association with self-reported of neurons and neural systems that regulate cognition,
symptoms of depression mood and behavior [53]. These changes could lead to
Consistent with earlier findings, we found that women symptoms of depression through disturbances in sleep-
exposed to controlling behavior, physical- and sexual and wake activity, decreased interest in feeding and social-
violence by an intimate partner were more likely to re- izing with others [53]. Likewise, it is widely known that
port symptoms of depression [16, 20, 22] compared to among people with a history of depression, previous expos-
women not exposed. A study performed on middle-aged ure to stressful life events is more common compared to
women in Australia for instance, found that of those those with no such history [20, 54].
‘sometimes’ or ‘often’ experiencing symptoms of depres- In line with previous research [55, 56] findings from the
sion, 20% and 28% respectively reported exposure to IPV current study revealed that women exposed to controlling
[27]. Traumatic and psychological stress reactions are con- behavior alone, i.e. without physical or sexual violence,
sidered to be the core mechanisms that explain why IPV during the past 12 months had higher odds to report
may cause subsequent depression in women [2]. Recent depressive-related symptoms compared to unexposed
Lövestad et al. BMC Public Health (2017) 17:335 Page 8 of 11

Table 5 Associations between partner violence exposure, covariates and symptoms of depression. Presented as adjusted Odds
Ratios and 95% confidence intervals (95% CI). N = 573 women
Explanatory variables and confounders Model 1 Model 2 Model 3
Controlling behavior past 12 months
(Unexposed vs. Exposed) 2.44 (1.61–3.71) 2.27 (1.48–3.50) 2.43 (1.56–3.79)
Age
(26–65 years vs 18–25) 0.85 (0.46–1.59) 0.90 (0.48–1.68) 0.83 (0.44–1.60)
Civil status
(Married, cohabitant vs. single, divorced) 2.51 (1.43–4.41) 2.62 (1.48–4.64) 2.75 (1.53–4.95)
Employment status
(Employed vs. not employed) 2.19 (1.42–3.40) 1.94 (1.23–3.06) 1.97 (1.23–3.14)
Social support
(Good vs. poor) 1.90 (1.27–2.86) 1.88 (1.24–2.85)
Grown up with- and witnessed IPV
(No/ unsure vs. witnessed IPV) 3.92 (1.72–8.91)
Physical violence past 12 months
(Unexposed vs. Exposed) 3.60 (1.84–7.04) 3.01 (1.52–5.98) 3.06 (1.50–6.24)
Age
(26–65 years vs 18–25) 0.86 (0.46–1.60) 0.86 (0.46–1.62) 0.81 (0.42–1.54)
Civil status
(Married, cohabitant vs. single, divorced) 2.61 (1.46–4.65) 2.73 (1.52–4.92) 2.80 (1.53–5.13)
Employment status
(Employed vs. not employed) 2.26 (1.46–3.50) 2.03 (1.28–3.21) 2.06 (1.30–3.29)
Social support
(Good vs. poor) 1.92 (1.28–2.88) 1.92 (1.27–2.90)
Grown up with- and witnessed IPV
(No/ unsure vs. witnessed IPV) 3.56 (1.58–8.02)
S
(Unexposed vs. Exposed) 5.02 (1.52–16.57) 4.47 (1.31–15.26) 4.67 (1.35–16.18)
Age
(26–65 years vs 18–25) 0.93 (0.50–1.73) 0.93 (0.49–1.75) 0.85 (0.44–1.64)
Civil status
(Married, cohabitant vs. single, divorced) 2.62 (1.48–4.64) 2.71 (1.51–4.88) 2.77 (1.52–5.01)
Employment status
(Employed vs. not employed) 2.13 (1.38–3.30) 1.91 (1.21–3.01) 1.96 (1.23–3.12)
Social support
(Good vs. poor) 2.09 (1.40–3.12) 2.08 (1.38–3.13)
Grown up with- and witnessed IPV
(No/ unsure vs. witnessed IPV) 3.53 (1.57–7.96)

women. This supports the repeatedly confirmed findings to employment and wage-earning [58]. Some authors sug-
in earlier literature that psychological abuse, including gest that within opposite-sex relationships, controlling be-
controlling behavior is as detrimental to women’s mental havior is perpetrated primarily by male partners [9, 58]
health as other forms of IPV [16, 57]. Controlling behavior and therefore frequently found to be experienced by
is used by the perpetrator in order to obtain obedience women from agency samples [9]. However, our findings
and dependency by depriving the partner from a range of together with previous research on population based sam-
important aspects in everyday life such as access to sup- ples indicate that controlling behavior is experienced also
port systems, economic resources, social life and the right among women in population-based samples [5, 18, 56]. By
Lövestad et al. BMC Public Health (2017) 17:335 Page 9 of 11

third-parties, controlling behavior is often perceived as in a female population based sample. However, some limi-
less harmful and more acceptable than physical violence tations should be considered when interpreting the results
[59]. Controlling behavior however, has shown to be at from this study. We are aware that women’s exposure to
least as harmful to women’s mental health as physical and psychological violence is not limited to acts of controlling
sexual violence, even in cases were no physical violence is behavior. It also includes other forms of psychological vio-
present [31]. It has likewise been demonstrated that when lence such as threats, intimidation and belittling [5] which
psychological violence includes power and control tactics, substantially contribute to women’s symptoms of depres-
the associations between psychological violence and de- sion [7]. However, our aim was to specifically address the
pressive symptoms further increase [56]. association between controlling behavior and symptoms
In accordance with previous studies, we found that of depression since this form of psychological abuse is
controlling behavior and physical violence were those qualitatively different compared to other forms of violence
forms of violence that overlapped to the largest extent in that it restricts women’s basic autonomy, liberty and
[16]. In our sample however, we found that 2.5% of the freedom [58] and may be more consequential for women’s
women were exposed to all three forms of violence mental health over time compared to other forms of
which is a smaller proportion compared to the findings verbal/ psychological abuse [6].
in a study by Thompson and his colleagues [14], where Questions about witnessing inter-parental violence as
30% of the women had experienced multiple forms of a child, IPV exposure and depressive symptoms were all
violence. Due to small sample size, further analyses of based on retrospective self-reporting and may therefore
associations between those exposed to physical or sexual have been subjected to recall bias, as well as to system-
violence only (n = 11 and n = 3 respectively) and self- atic non-disclosure. This in turn may have led to an
reported symptoms of depression was not possible. It is underestimation of reported associations. Past emotions
further important to note that since we did not adjust and behaviors are difficult to recall in an accurate way
for each type of IPV exposure when analyzing associa- and historical responses of psychiatric symptoms might
tions between IPV and symptoms of depression, it is be biased by the respondents’ current mental health
likely that there might be an interaction between- or an status [61]. Further, obtaining reliable data on IPV is a
additive effect of the different types of IPV and its asso- difficult and complex task since it cannot be directly ob-
ciation with symptoms of depression. served and in addition, it is surrounded by taboos, feel-
We found that poor social support was independently ings of guilt, fear and shame [62].
associated with self-reported symptoms of depression, Consistent with earlier findings in research on IPV
which is congruent with earlier research [21, 57]. Good [11], women born outside Sweden, of younger age, un-
social support has been shown to predict recovery from married and with a lower annual income were somewhat
depressive symptoms among women exposed to IPV underrepresented in our study population. Rates of IPV
[57], whereas an accumulation of poor social support at exposure are found to be higher in these groups [3, 36]
a younger age is independently associated with internal- hence the prevalence of exposure to IPV in this study
izing symptoms later in life [60]. This suggests that im- might be underestimated and consequently have led to an
proving women’s access to social support would mitigate underestimation of its associations with depressive related
symptoms of depression in women exposed to IPV. symptoms. According to previous research, symptoms of
Consistent with previous research, we also found that depression might also be underestimated in our study,
women being single, widowed or divorced [21], being since sociodemographic factors such as living alone, low
unemployed [33] and having grown up with- and wit- income and younger age are related both to mental disor-
nessed inter-parental violence [34] contributed to self- ders and non- participation rates in studies [63]. Since this
reported symptoms of depression. Previous research has study was a cross sectional survey, we did not have any in-
shown that unemployed women are at increased risk for formation on the temporal relationship between the onset
depressive symptoms compared to employed women, and end of the reported violence or the depressive symp-
and the risk increases even more when women are ex- toms. This precludes the determination of causality, i.e.
posed to IPV [33]. This together with our findings sug- IPV may cause depression but depression may cause IPV.
gests that social support and restricted occupational However, findings from longitudinal studies suggest a con-
opportunities may interact with the relationship between sistent and independent causal link between exposure to
IPV and mental health problems. IPV and depressive symptoms [21, 64].
Due to few cases in each of the frequency categories, the
Methodological considerations included frequency levels could not be investigated as sin-
A strength with this study is that few if any earlier studies gle entities. Co-occurrence of different forms of IPV and
performed in Sweden, have analyzed the associations be- its association with symptoms of depression gave small
tween controlling behavior and symptoms of depression sample sizes and no further analyses were performed.
Lövestad et al. BMC Public Health (2017) 17:335 Page 10 of 11

Another limitation is that we did not use any recog- 2

Department of Clinical Neuroscience, Division of Insurance Medicine,
nized and validated instrument to measure self-reported Karolinska Institutet, SE- 171 77 Stockholm, Sweden.

symptoms of depression. However, from a clinical point Received: 9 February 2017 Accepted: 1 April 2017
of view, many women present at the health care center
with symptoms of depression and therefore it is important
to make clinicians aware of the fact that such symptoms
References
may mirror depression that eventually may be associated 1. Ellsberg M, Jansen HA, Heise L, Watts CH, Garcia-Moreno C. Intimate partner
with exposure to IPV. violence and women's physical and mental health in the WHO multi-
country study on women's health and domestic violence: an observational
study. Lancet. 2008;371(9619):1165–72.
Conclusions 2. World Health Organization: Global and regional estimates of violence
A strength with this study is the analysis of controlling against women: prevalence and health effects of intimate partner violence
and nonpartner sexual violence. Geneva: World Health Organization; 2013.
behavior and its association with self-reported symptoms
3. World Health Organization. Preventing intimate partner and sexual violence
of depression. Findings from the current study supports against women: taking action and generating evidence. Geneva: World
previous research in that women’s exposure to control- Health Organization; 2010.
4. World Health Organization. Summary report. WHO Multi-country Study on
ling behavior, as well as physical and sexual violence per-
Women’s Health and Domestic Violence againstWomen. Initial results on
petrated by a male intimate partner is clearly associated prevalence, health outcomes and women’s responses. Geneva: World
with symptoms of depression. Health Organization; 2005.
5. European Union Agency for Fundamental Rights. Violence against women:
Abbreviations an EU-wide survey. Main results. Luxembourg: European Union Agency for
CI: Confidence Interval; CSB: Controlling Behavior Scale; DSM-IV: Statistical Fundamental Rights; 2014.
Manual of Mental Disorders, Fourth Edition; IPV: Intimate Partner Violence; 6. Katz J, Arias I. Psychological abuse and depressive symptoms in dating
LNU: Swedish Level of Living Survey; OR: Odds Ratio; WHO: World Health women: do different types of abuse have differential effects? J Fam
Organization Violence. 1999;14(3):281–95.
7. Mechanic MB, Weaver TL, Resick PA. Mental health consequences of
Acknowledgements intimate partner abuse: a multidimensional assessment of four different
The authors would like to thank Chris Pickering who kindly made comments forms of abuse. Violence Against Women. 2008;14(6):634–54.
and suggestions for the improvement of this paper. 8. Chang JC, Cluss PA, Burke JG, Hawker L, Dado D, Goldstrohm S, Scholle SH.
Partner violence screening in mental health. Gen Hosp Psychiatry. 2011;
Funding 33(1):58–65.
This study was supported by a grant from the Swedish Research Council for 9. Johnson MP. Conflict and control: gender symmetry and asymmetry in
Health, Working Life and Welfare (Forte). domestic violence. Violence Against Women. 2006;12(11):1003–18.
10. Lovestad S, Krantz G. Men's and women's exposure and perpetration of
Availability of data and materials partner violence: an epidemiological study from Sweden. BMC Public
The data used for this study is stored at the Department of Public Health Health. 2012;12:945.
and Community Medicine at the University of Gothenburg. For access, 11. Andersson, T, Heimer G, Lucas S. Violence and health in Sweden - A
please contact the corresponding author. National Prevalence Study on Exposure to Violence among Women and
Men and its Association to Health. Uppsala: National Centre for Knowledge
Authors’ contributions on Men’s Violence Against Women; 2014.
SL carried out all statistical analyses, wrote the first draft of the manuscript 12. DLL KEG, Mercy James A, Zwi Anthony B, Rafael L. The world report on
and revised new drafts based on input from co-authors. GK is the main violence and health. Geneva: World Health Organization (WHO); 2002.
responsible researcher for this research, designed the programme, and was 13. Carlson BE, McNutt L-A, Choi DY. Childhood and adult abuse among
responsible for developing the questionnaire and directed the data collection. women in primary health care: effects on mental health. J Interpersonal
GK with SL, JL and MV designed this project, planned the analyses and gave Violence. 2003;18(8):924–41.
input on manuscript drafts. GK, JL and MV read and revised the text until a final 14. Thompson RS, Bonomi AE, Anderson M, Reid RJ, Dimer JA, Carrell D, Rivara
version was in place. All authors read and approved the final manuscript. FP. Intimate partner violence: prevalence, types, and chronicity in adult
women. Am J Prev Med. 2006;30(6):447–57.
Competing interests 15. Montero I, Martin-Baena D, Escriba-Aguir V, Ruiz-Perez I, Vives-Cases C,
The authors declare that they have no competing interests. Talavera M. Intimate partner violence in older women in Spain: prevalence,
health consequences, and service utilization. J Women Aging. 2013;25(4):358–71.
Consent for publication 16. Pico-Alfonso MA, Garcia-Linares MI, Celda-Navarro N, Blasco-Ros C,
Not applicable. Echeburua E, Martinez M. The impact of physical, psychological, and sexual
intimate male partner violence on women's mental health: depressive
Ethical approval and consent to participate symptoms, posttraumatic stress disorder, state anxiety, and suicide.
Ethical approval was provided from the Regional Ethics Review Board in J Women's Health. 2006;15(5):599–611.
Gothenburg (Dnr: 527–08). A short letter with information about the study 17. Renner LM, Habib L, Stromquist AM, Peek-Asa C. The association of intimate
and possibility to deny further participation was sent out to all selected partner violence and depressive symptoms in a cohort of rural couples.
individuals. J Rural Health. 2014;30(1):50–8.
18. Lundgren E, Heimer G, Westerstrand J, Kalliokoski A-M. Slagen Dam: Mäns
Våldmot kvinnor i jämställda Sverige-enomfångs undersökning (Captured
Publisher’s note queen - men’s violence against women in "equal" Sweden - a prevalence
Springer Nature remains neutral with regard to jurisdictional claims in published study). Umeå: Brottsoffermyndigheten; 2001.
maps and institutional affiliations. 19. Dillon G, Hussain R, Loxton D, Rahman S. Mental and physical health and
intimate partner violence against women: a review of the literature. Int J
Author details Family Med. 2013;2013:313909.
1
Department of Community Medicine and Public Health, Sahlgrenska 20. Beydoun HA, Beydoun MA, Kaufman JS, Lo B, Zonderman AB. Intimate
Academy at University of Gothenburg, Box 453, 405 30 Göteborg, Sweden. partner violence against adult women and its association with major
Lövestad et al. BMC Public Health (2017) 17:335 Page 11 of 11

depressive disorder, depressive symptoms and postpartum depression: 44. Samelius L, Wijma B, Wingren G, Wijma K. Lifetime history of abuse,
a systematic review and meta-analysis. Soc Sci Med. 2012;75(6):959–75. suffering and psychological health. Nordic J Psychiatr. 2010;64(4):227–32.
21. Chuang CH, Cattoi AL, McCall-Hosenfeld JS, Camacho F, Dyer AM, Weisman 45. Dufort M, Stenbacka M, Gumpert CH. Physical domestic violence exposure is
CS. Longitudinal association of intimate partner violence and depressive highly associated with suicidal attempts in both women and men. Results from
symptoms. Ment Health Fam Med. 2012;9(2):107–14. the national public health survey in Sweden. Eur J Public Health 2015, 25(3):413–8.
22. Devries KM, Mak JY, Bacchus LJ, Child JC, Falder G, Petzold M, Astbury J, 46. Nybergh L, Taft C, Enander V, Krantz G. Self-reported exposure to intimate
Watts CH. Intimate partner violence and incident depressive symptoms and partner violence among women and men in Sweden: results from a
suicide attempts: a systematic review of longitudinal studies. PLoS Med. population-based survey. BMC Public Health. 2013;13:845.
2013;10(5):e1001439. 47. American Psychiatric Association. Diagnostic and statistical manual of mental
23. World health Organization DoMHaSD: Gender disparities in mental health. disorders: DSM-IV-TR. Washington: American Psychiatric Association; 2000.
http://www.who.int/mental_health/media/en/242.pdf. Accessed 8 Apr 2017. 48. Garcia-Moreno C, Jansen H, Ellsberg M, Heise L, Watts C. Multi-country
24. Western D, SpringerLink. Gender-based violence and depression in women: Study on Women's Health and Domestic Violence against women. Initial
a feminist group work response / [Elektronisk resurs]. Springer New York: results on prevalence, health outcomes and women's responses. Geneva:
New York; 2013. World Health Organization; 2005.
25. Sobocki P, Lekander I, Borgström F, Ström O, Runeson B. The economic burden 49. Nybergh L, Taft C, Krantz G: Psychometric properties of the WHO Violence
of depression in Sweden from 1997 to 2005. Eur Psychol. 2007;22(3):146–52. Against Women instrument in a female population-based sample in
26. Filson J, Ulloa E, Runfola C, Hokoda A. Does powerlessness explain the Sweden: a cross-sectional survey. BMJ Open 2013, 3(5).
relationship between intimate partner violence and depression? 50. Graham-Kevan N, Archer J. Intimate terrorism and common couple violence.
J Interpersonal Violence. 2010;25(3):400–15. A test of Johnson's predictions in four British samples. J Interpersonal
27. Loxton D, Schofield M, Hussain R. Psychological health in midlife among Violence. 2003;18(11):1247–70.
women who have ever lived with a violent partner or spouse. 51. Östberg V, Lennartsson C. Getting by with a little help: the importance of
J Interpersonal Violence. 2006;21(8):1092–107. various types of social support for health problems. Scand J Public Health.
28. Cavanaugh CE, Messing JT, Petras H, Fowler B, La Flair L, Kub J, Agnew J, 2007;35(2):197–204.
Fitzgerald S, Bolyard R, Campbell JC. Patterns of violence against women: a 52. World Health Organization: Putting women first: Ethical and Safety
latent class analysis. Psychol Trauma. 2012;4(2):169–76. Recommendations for Research on Domestic Violence Against Women.
29. Johnson WL, Giordano PC, Longmore MA, Manning WD. Intimate partner Geneva: World Health Organization; 2001.
violence and depressive symptoms during adolescence and young 53. Slavich GM, Irwin MR. From stress to inflammation and major depressive
adulthood. J Health Soc Behav. 2014;55(1):39–55. disorder: a social signal transduction theory of depression. Psychol Bull.
30. Mburia-Mwalili A, Clements-Nolle K, Lee W, Shadley M, Wei Y. Intimate 2014;140(3):774–815.
partner violence and depression in a population-based sample of women: 54. Kessler RC. Epidemiology of women and depression. J Affect Disord. 2003;
can social support help? J Interpersonal Violence. 2010;25(12):2258–78. 74(1):5–13.
31. Anderson KL. Is partner violence worse in the context of control? J Marital 55. Hellemans S, Buysse A, De Smet O, Wietzker A. Intimate partner violence in
Fam Ther. 2008;70(5):1157–68. Belgium: prevalence, individual health outcomes, and relational correlates.
32. Rich JL, Byrne JM, Curryer C, Byles JE, Loxton D. Prevalence and correlates of Psychol Belg. 2014;54(1):79–96.
depression among Australian women: a systematic literature review, January 56. Coker AL, Davis KE, Arias I, Desai S, Sanderson M, Brandt HM, Smith PH.
1999- January 2010. BMC Res Notes. 2013;6:424. Physical and mental health effects of intimate partner violence for men and
33. Dougé N, Lehman EB, McCall-Hosenfeld JS. Social support and employment women. Am J Prev Med. 2002;23(4):260–8.
status modify the effect of intimate partner violence on depression 57. Blasco-Ros C, Sanchez-Lorente S, Martinez M. Recovery from depressive
symptom severity in women: results from the 2006 behavioral risk factor symptoms, state anxiety and post-traumatic stress disorder in women
surveillance system survey. Womens Health Issues. 2014;24(4):e425–34. exposed to physical and psychological, but not to psychological intimate
34. Iverson KM, McLaughlin KA, Gerber MR, Dick A, Smith BN, Bell ME, Cook N, partner violence alone: a longitudinal study. BMC Psychiatr. 2010;10:98.
Mitchell KS. Exposure to interpersonal violence and its associations with 58. Stark E. Looking beyond domestic violence: policing coercive control.
psychiatric morbidity in a U.S. National Sample: a gender comparison. J Police Crisis Negotiations. 2012;12(2):199–217.
Psychol Violence. 2013;3(3):273–87. 59. Williams C, Richardson DS, Hammock GS, Janit AS. Perceptions of physical
and psychological aggression in close relationships: a review. Aggress
35. Bengtsson-Tops A, Markström U, Lewin B. The prevalence of abuse in
Violent Behav. 2012;17(6):489–94.
Swedish female psychiatric users, the perpetrators and places where abuse
60. Landstedt E, Gustafsson PE, Johansson K, Hammarstrom A. Longitudinal
occurred. Nordic J Psychiatr. 2005;59(6):504–10.
associations between social relationships at age 30 and internalising
36. Finnbogadottir H, Dykes AK, Wann-Hansson C. Prevalence of domestic
symptoms at age 42: findings from the northern Swedish cohort. Int J
violence during pregnancy and related risk factors: a cross-sectional study in
Public Health. 2016;61(1):75–81.
southern Sweden. BMC Womens Health. 2014;14:63.
61. Tsuang MT, Tohen M, Jones PB. Textbook in psychiatric epidemiology. third
37. Hedin LW, Janson PO. The invisible wounds: the occurrence of
ed. Hoboken: Wiley-Blackwell; 2011.
psychological abuse and anxiety compared with previous experience of
62. Ruiz-Perez I, Plazaola-Castano J, Vives-Cases C. Methodological issues in the
physical abuse during the childbearing year. J Psychosom Obstet Gynaecol.
study of violence against women. J Epidemiol Community Health. 2007;
1999;20(3):136–44.
61(Suppl 2):ii26–31.
38. Radestad I, Rubertsson C, Ebeling M, Hildingsson I. What factors in early
63. Lundberg I, Damstrom Thakker K, Hallstrom T, Forsell Y. Determinants of
pregnancy indicate that the mother will be hit by her partner during the
non-participation, and the effects of non-participation on potential cause-
year after childbirth? A nationwide Swedish survey. Birth. 2004;31(2):84–92.
effect relationships, in the PART study on mental disorders. Soc Psychiatry
39. Bergman B, Brismar B. A 5-year follow-up study of 117 battered women. Am
Psychiatr Epidemiol. 2005;40(6):475–83.
J Public Health. 1991;81(11):1486–9.
64. La Flair LN, Bradshaw CP, Campbell JC. Intimate partner violence/abuse and
40. Bengtsson-Tops A, Tops D. Self-reported consequences and needs for
depressive symptoms among female health care workers: longitudinal
support associated with abuse in female users of psychiatric care. Int J Ment findings. Women’s Health Issues. 2012;22(1):e53–9.
Health Nurs. 2007;16(1):35–43.
41. Fernbrant C, Emmelin M, Essen B, Ostergren PO, Cantor-Graae E. Intimate
partner violence and poor mental health among Thai women residing in
Sweden. Glob Health Action. 2014;7:24991.
42. Dufort M, Gumpert CH, Stenbacka M. Intimate partner violence and help-
seeking–a cross-sectional study of women in Sweden. BMC Public Health.
2013;13:866.
43. Krantz G, Ostergren PO. The association between violence victimisation and
common symptoms in Swedish women. J Epidemiol Community Health.
2000;54(11):815–21.