Gene Feder and Louise M. Howard are members of the WHO Guideline Development Group on Policy and Practice Guidelines for responding to Violence Against Women and the NICE/SCIE Guideline Development Group on Preventing and Reducing Domestic Violence. KT and SO declare no competing interests. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: LMH GF SO KT. Performed the experiments: KT SO. Analyzed the data: KT SO. Contributed reagents/materials/analysis tools: KT SO. Wrote the paper: KT SO GF LMH. Critical revisions of the manuscript: KT SO GF LMH.
Little is known about the extent to which being a victim of domestic violence is associated with different mental disorders in men and women. We aimed to estimate the prevalence and odds of being a victim of domestic violence by diagnostic category and sex.
Study design: Systematic review and meta-analysis. Data Sources: Eighteen biomedical and social sciences databases (including MEDLINE, EMBASE, PsycINFO); journal hand searches; scrutiny of references and citation tracking of included articles; expert recommendations, and an update of a systematic review on victimisation and mental disorder. Inclusion criteria: observational and intervention studies reporting prevalence or odds of being a victim of domestic violence in men and women (aged ≥16 years), using validated diagnostic measures of mental disorder. Procedure: Data were extracted and study quality independently appraised by two reviewers. Analysis: Random effects meta-analyses were used to pool estimates of prevalence and odds.
Forty-one studies were included. There is a higher risk of experiencing adult lifetime partner violence among women with depressive disorders (OR 2.77 (95% CI 1.96–3.92), anxiety disorders (OR 4.08 (95% CI 2.39–6.97), and PTSD (OR 7.34 95% CI 4.50–11.98), compared to women without mental disorders. Insufficient data were available to calculate pooled odds for other mental disorders, family violence (i.e. violence perpetrated by a non-partner), or violence experienced by men. Individual studies reported increased odds for women and men for all diagnostic categories, including psychoses, with a higher prevalence reported for women. Few longitudinal studies were found so the direction of causality could not be investigated.
There is a high prevalence and increased likelihood of being a victim of domestic violence in men and women across all diagnostic categories, compared to people without disorders. Longitudinal studies are needed to identify pathways to being a victim of domestic violence to optimise healthcare responses.
Domestic violence is an international public health problem, affecting the lives of hundreds of thousands of people every year. Globally, prevalence estimates of lifetime experiences of physical or sexual partner violence among women range from 15%–71%, with past year estimates ranging from 4% and 54%
As a consequence of the substantial physical and psychiatric morbidity associated with domestic violence
Prolonged exposure to threatening life events, including domestic violence, is associated with the onset, duration and recurrence of mental disorders
The prevalence (lifetime and past year) of being a victim of domestic violence in men and women with mental disorders
The odds of being a victim of domestic violence in men and women with mental disorders compared with non-mentally disordered controls
This review followed MOOSE and PRISMA guidelines
Studies were eligible for inclusion if they: (a) included men and/or women who were 16 years or older and were diagnosed with a mental disorder using a validated diagnostic instrument (i.e. diagnostic instruments that have been validated against a gold standard measure for diagnosing mental disorder, such as the Schedules for Clinical Assessment in Neuropsychiatry
The downloaded titles and abstracts were screened against the inclusion criteria by two reviewers (KT and SO). If it was unclear whether a reference met the inclusion criteria, it was taken forward to the next stage of screening. The full texts of potentially eligible studies were assessed by two reviewers (KT and SO). If studies collected data on the prevalence and/or odds of domestic violence but did not report it, authors were contacted for the data. Details of the 1,083 excluded papers and reasons for exclusion are available upon request.
Data from included papers were extracted into a standardised electronic database by two reviewers (KT and SO) and a random sample of 20% was independently cross-checked. Extracted data included details on: the study design; sample characteristics; measures of mental disorder and domestic violence, and the prevalence and odds of lifetime/past year domestic violence. Data were extracted separately for men and women. When reported, details on resource use, impact and severity of violence and chronicity of mental disorders were extracted.
The quality of included studies was independently appraised by two reviewers (KT and SO) using criteria adapted from validated tools
We calculated prevalence, odds ratios and 95% confidence intervals for domestic violence among men and women by type of mental disorder. If studies measured multiple disorders, odds ratios were calculated separately by type of mental disorder and for each estimate the control group were participants without any mental disorder. Prevalence and odds ratios were also calculated separately by sex and type of violence. Regarding type of violence, we report results for any violence (i.e. physical, sexual and psychological violence) and for physical violence alone. Data on the prevalence and odds of sexual and psychological violence were limited and are given in
We calculated DerSimonian-Laird random effects odds ratio estimates (with corresponding 95% confidence intervals) for lifetime and past year domestic violence among people with mental disorders, compared to people without a mental disorder, if reports were available from three or more high-quality studies
The study selection process is presented in
Key characteristics of the studies are reported in
Total (n = 41) | Lifetime domestic violence (n = 26) | Past year domestic violence (n = 18) | |
Sample: | |||
Males only | 0 | 0 | 0 |
Females only | 25 | 14 | 14 |
Males and females | 16 |
12 | 4 |
Diagnoses: | |||
Schizophrenia & non-affective psychosis | 3 | 2 | 1 |
Bipolar affective disorder | 2 | 2 | 0 |
Depressive disorders | 26 | 21 | 12 |
Dysthymia | 5 | 2 | 3 |
Anxiety disorders | 15 | 9 | 7 |
PTSD | 14 | 9 | 7 |
OCD | 2 | 2 | 0 |
Panic disorders | 6 | 2 | 4 |
Phobias | 3 | 2 | 1 |
Somatisation | 1 | 0 | 1 |
Eating disorder | 1 | 0 | 1 |
Personality disorder | 4 | 3 | 1 |
Common Mental Disorder | 5 | 4 | 2 |
Setting: | |||
Clinical | 17 | 11 | 7 |
Non-clinical | 24 | 15 | 11 |
Perpetrator: | |||
Partner only | 38 |
24 | 17 |
Family only | 0 | 0 | 0 |
Partner or family | 3 | 2 | 1 |
Type of violence | |||
Physical violence | 20 | 15 | 5 |
Psychological violence | 9 | 5 | 4 |
Sexual violence | 4 | 3 | 1 |
Physical, sexual, psychological combined | 11 | 5 | 6 |
Recency of violence | |||
Lifetime domestic violence | 23 | ||
Past year domestic violence | 15 | - | - |
Lifetime and past year domestic violence | 3 | ||
Measurement of domestic violence | |||
Validated measures | 18 |
11 | 8 |
Non-validated measures | 19 |
11 | 7 |
Trauma items from DSM/CIDI criteria | 4 | 4 | 0 |
Region: | |||
North America | 17 | 10 | 9 |
Central America | 1 | 1 | 0 |
South America | 1 | 1 | 0 |
Europe | 6 | 5 | 2 |
Africa | 3 | 3 | 0 |
Asia | 8 | 6 | 2 |
Australasia | 5 | 0 | 5 |
Categories are not mutually exclusive and rows may therefore add to >40.
Sex-disaggregated data was available for 11 of the 16 studies.
Five papers measured only spousal violence.
Four papers made modifications to validated measures and five did not use all items in the measure.
In 16 studies the authors developed their own measure to assess domestic violence.
Results present data for lifetime and past year experiences of any type of partner violence (i.e. physical, sexual and/or psychological violence) across all mental disorders in women and men. Prevalence and odds estimates for all included studies are presented in
Among women, the median prevalence of any lifetime partner violence (7 studies) was 45.8% (IQR 21.3%–76.5%; range 15.6%–89.2%)
The median prevalence of any past year partner violence (7 studies) was 35.3% (IQR 16.0%–40.1%; range 1.7%–82.5%) among women with depressive disorders
Two high-quality studies reported on the prevalence of any lifetime partner violence among men and women with dysthymia
A cross-sectional survey of 364 pregnant and postpartum Vietnamese women reported a prevalence of 16.7% for any past year partner violence among six women with dysthymic disorder
The median prevalence of any lifetime partner violence (5 studies) for women with anxiety disorders was 27.6% (IQR 24.9%–72.7%; range 22.4%–89.9%)
Among women with anxiety disorders, the median prevalence of any past year partner violence (4 studies) was 28.4% (IQR 25.5%–42.2%, range 20.0%–80.5%)
The median prevalence of any lifetime partner violence (4 studies) among women with PTSD was 61.0% (IQR 41.1%–80.1%; range 29.4%–89.5%)
A survey of female welfare recipients found that 27.0% of women with PTSD had experienced physical partner violence in the past year. The study reported a greater likelihood of past year physical partner violence among women with PTSD compared to women without a mental disorder (OR 3.62; 95% CI 2.32–5.67)
A cross-sectional survey of 650 women attending primary care clinics in Pakistan reported a prevalence of 93.8% for any lifetime partner violence among women with obsessive compulsive disorder (OCD). The study found that women with OCD had an increased likelihood of experiencing any lifetime partner violence compared to women without a mental disorder (OR 6.43; 95% CI 1.95–33.23)
No high-quality studies reported the prevalence or odds of past year domestic violence among men or women with OCD.
No high-quality studies reported the prevalence or odds of lifetime domestic violence among men or women with an eating disorder.
One birth cohort study reported that at age 21, 63.6% of 11 women with eating disorders reported past year physical partner violence; women with eating disorders were more likely to report partner violence compared to women without a mental disorder (OR 7.31 95% CI 1.76–35.10)
One study, a national survey of 34,653 non-institutionalised American residents, reported an increased odds of lifetime physical partner violence among both women (OR: 6.06 95% CI 5.35–6.86) and men (OR: 7.04 95% CI 5.30–9.43) with any personality disorder, and lifetime prevalence estimates of 21.4% and 5.4% respectively
One birth cohort study reported that at age 21, 100.0% of three women with an antisocial personality disorder reported past year physical partner violence
The median prevalence of any lifetime partner violence among women with CMDs (3 studies) was 48.0% (IQR 35.6%–63.2%, range 23.0%–78.1%); women with CMDs were reported to be more likely to experience any lifetime partner violence compared to those without a mental disorder
A national survey of 7,047 UK householders reported an increased odds of any past year partner violence among women (OR: 4.4 95% CI: 3.32–5.82) and men (OR: 3.1 95% CI 2.18–4.39) with CMDs; prevalence estimates were 15.2% and 11.7% respectively
No high-quality studies reported the prevalence or odds of any lifetime partner violence among men or women with schizophrenia and non-affective psychosis. Two lower-quality studies, both conducted with psychiatric samples, reported that the lifetime prevalence of any partner violence ranged from 43.8%–83.3% among women with schizophrenia and non-affective psychosis
One birth cohort study reported a prevalence of 43.8% for past year physical partner violence among 16 women with non-affective psychosis
One study, a nationally representative survey of 34,563 non-institutionalised American residents, identified an increased odds of lifetime physical partner violence among both women (OR 8.14; 95% CI 6.99–9.47) and men (OR 9.42; 95% CI 6.57–13.50) with bipolar disorder, and lifetime prevalence estimates of 26.7% and 7.1%, respectively
No high-quality studies reported the prevalence or odds of past year domestic violence among men or women with bipolar disorder.
Three studies presented longitudinal data on the relationship between mental disorders and domestic violence
Limited data were available on the impact of experiences of domestic violence and victims' resource use. Two studies reported increased odds of substance misuse problems (OR of 3.4 and 4.1) among people experiencing domestic violence
We found consistent evidence that both men and women with all types of mental disorders report a high prevalence and increased odds of domestic violence compared to people without mental disorder, with women more likely to experience abuse than men. Due to the limited number of high-quality studies it was not possible to calculate pooled odds of partner violence among men or for men or women with disorders other than depression, anxiety or PTSD. Studies on the prevalence and odds of domestic violence by non-intimate family members were also limited. Nonetheless, across a range of diagnoses, studies indicated that men and women with a mental disorder are at an increased likelihood of experiencing domestic violence compared to those without a mental disorder. For example, data from Wave II of the large US National Epidemiologic Survey on Alcohol and Related Conditions suggests that men and women with bipolar affective disorder were more than eight times more likely to report ever having been a victim of partner violence than people with no mental disorder
Although a bi-directional causal relationship between domestic violence and mental disorder seems likely
We used an inclusive search strategy and followed MOOSE and PRISMA reporting guidelines
All pooled odds ratio estimates indicated that women with mental disorders are at an increased likelihood of experiencing partner violence compared to women without mental disorders. However, in light of the high heterogeneity observed between studies, caution should be exercised when interpreting these figures. Due to a lack of data it was not possible to control for confounding factors when pooling prevalence estimates. When we excluded studies that used conservative definitions of domestic violence or employed non-validated instruments to measure domestic violence, heterogeneity was reduced. However, we do not know the relative contributions of the study setting and measurement of domestic violence to the heterogeneity, and it is likely that study country and known confounding factors (e.g. age, experiences of childhood abuse and substance misuse) may also affect variations in prevalence estimates. Funnel plot asymmetry also indicated the potential for publication bias among studies of depression.
Due to the lack of consistency in the data collected by the primary studies, we were unable to adjust our pooled estimates for potential confounders (e.g. childhood abuse). Furthermore, because of a lack of primary studies, we were unable to: calculate pooled estimates of the odds of domestic violence among men with mental disorders; to assess whether the odds of violence perpetrated by family members was increased among men and women with mental disorder; to analyse whether the prevalence and odds of domestic violence among men and women with mental disorder varied according to sexual preference.
Our meta-analyses were constrained by methodological and conceptual weaknesses in the primary studies. A third of studies scored <50% on quality appraisal criteria relating to selection bias; 23 studies used non-probability sampling, 15 did not provide information on the representativeness of their samples and 14 did not report on the likely impact of non-participation. Although most studies did not score poorly in relation to measurement bias, the measurement of domestic violence varied substantially, with regards to time period (lifetime vs. past year), type of abusive behaviour (physical, sexual, psychological or a combination of behaviours), and instrument. We reported separate estimates of the prevalence and odds of lifetime and past year domestic violence, but recognise that both measures are potentially problematic: recall bias may be present in studies that measure lifetime domestic violence, while participants in studies of past year violence may have had insufficient time to acknowledge or identify their abuse experiences as such
This systematic review provides strong evidence of a high prevalence and increased odds of domestic violence across all mental disorders among both men and women and draws attention to key gaps in the evidence base. The findings of this review highlight the need for healthcare professionals to recognise the increased vulnerability of men and women with mental disorders to domestic violence and to be prepared to identify and address these issues in treatment plans. Current evidence suggests that identification of domestic violence is most effective when professionals are trained to understand the nature of domestic violence and its long term impact on health, to ask about domestic violence safely if abuse is ongoing, and have clear referral and care pathways for identified victims
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(DOC)
(TIF)
(TIF)
(TIF)
We would like to thank the experts in this area who responded so helpfully in our request for data. We also gratefully acknowledge Fraser Anderson for her assistance during summer 2011; Mauricio Moreno for technical support; Professor Michael Dewey and the Institute of Psychiatry Biostatistics team for statistical support, and Seynam Kluvitse for administrative assistance. We also wish to thank Monika Janosik, Leo Koeser, Sharron Leung, Kazuyo Machiyama, Mauricio Moreno, Michael Dewey, Oliver Schauman, and Tracy Teng for translating articles that were published in non-English languages.