Domestic violence, lifetime trauma and psychological health of childbearing women


Dr G. Mezey, Department of Mental Health (Forensic), St George's Hospital Medical School, Jenner Wing, Ground Floor, Cranmer Terrace, London, UK.


Objective  Although histories of abuse are associated with psychiatric illness in women, health professionals rarely enquire directly about such experiences. This study examined the association between physical and sexual violence and lifetime trauma and depressive and posttraumatic stress symptoms in women receiving maternity care.

Design  Cross sectional study.

Setting  South London Hospital maternity services.

Population  Two hundred women receiving postnatal or antenatal care.

Methods  Two hundred women receiving postnatal or antenatal care at a South London maternity service were screened for lifetime experiences of trauma and domestic violence. Information was obtained about self-harming behaviour, suicidal thoughts and attempts and psychiatric history. Women completed the Edinburgh Postnatal Depression Scale (EPDS) and the Posttraumatic Diagnostic Scale (PTDS).

Main outcome measures  Results  One hundred and twenty-one (60.5%) women reported at least one traumatic event and two-thirds of these had experienced multiple traumatic events. The most frequent (34%) was witnessing or experiencing physical assault by a family member. Forty-seven (23.5%) women had experienced domestic violence. Physical and sexual abuse commonly co-occurred. Thirteen (10.7%) women with a trauma history had current posttraumatic stress disorder. Severe posttraumatic symptoms were associated with physical and sexual abuse histories and repeat victimisation. Adult and childhood physical and sexual abuse histories were also associated with more severe depressive symptomatology. Significant social factors associated with depression were being single, separated or in a non-cohabiting relationship.

Conclusion  Traumatic events are under-recognised risk factors in the development of depressive and posttraumatic stress symptoms in childbearing women. Childhood abuse creates a vulnerability to re-traumatisation in adulthood. Awareness of the impact of trauma and abuse on psychological health may enable more appropriate targeting of clinical services and support for women receiving maternity care.


It is recognised that severe or life-threatening trauma is a risk factor for the development of psychiatric illness including posttraumatic stress disorder and depression.1 According to the Diagnostic and Statistical Manual for Mental Disorders (DSM-IV), a traumatic event is defined as one ‘that involves actual or threatened death, serious injury or threat to physical integrity’ and gives rise to feelings of intense fear, horror or helplessness.2 The risk of developing posttraumatic stress disorder following a traumatic event depends on a number of factors, including the type of trauma, and is highest following assaultative violence.3 Although women are at less risk than men of being exposed to major or life-threatening trauma, once exposed they are twice as likely to develop posttraumatic stress disorder.4 Women also experience different kinds of traumatic events to men; with higher rates of sexual and domestic violence, which are associated with the highest rates of posttraumatic stress disorder.5 Domestic violence is under-reported and under-recognised across a range of health settings.6 Health professionals rarely enquire about domestic violence and women are reluctant to disclose such experiences in the absence of direct questioning.7,8

A number of studies have found that, once exposed to a traumatic event, the risk of experiencing a subsequent trauma is substantially increased.9 Women who have been sexually abused as children are significantly more likely to report subsequent abuse as an adult, including domestic violence, than women without a childhood abuse history.10 It is possible that domestic violence acts as a marker for other traumatic and adverse life events, which mediate the psychiatric and social difficulties that are often attributed to domestic violence. However, what is not clear, is whether the risk of re-victimisation relates exclusively to physical and sexual violence or whether it is increased for all types of trauma, outside the context of interpersonal violence.

Studies of depression during pregnancy and the postpartum period have generally failed to assess for histories of abuse and violence as potential risk factors,11,12 in spite of the suggestion that women may be particularly vulnerable to domestic violence during this time.13

This study was part of a larger ESRC funded research project on domestic violence in pregnancy, which was carried out in the maternity services of two South London teaching hospitals. The aim of this study was to examine the prevalence and types of traumatic events, reported by women and the impact of the various types of traumatic events on women's mental health, particularly with respect to posttraumatic stress disorder and depression. The study received ethical approval from St Thomas' Hospital Research Ethics Committee.


The study used a cross sectional design that consisted of 200 English-speaking women, aged 16 and older, from the postnatal and antenatal wards and the Day Assessment Unit (antenatal) at Guy's and St Thomas' maternity service between July 2001 and April 2002.

As the primary aim of this study was to estimate the prevalence of domestic violence in an obstetric population, the study was powered to be able to estimate the prevalence within an acceptable error margin. Therefore, it was planned to recruit 200 patients to be able to calculate the 95% confidence interval around the estimate of prevalence of domestic violence to within ±5%.

Women on the postnatal wards were generally discharged within 12 hours of giving birth, unless there were complications with the mother or the baby, or they had an instrumental delivery or caesarean section. Women on the antenatal ward were admitted as a result of an obstetric complication and could remain on the ward for anything between a few days to a few months. The Day Assessment Unit served as a walk-in clinic for women of <18 weeks of gestation, who were experiencing problems and required an assessment. The women attending this clinic could be discharged within the day, or admitted to the antenatal ward. All women on the antenatal ward were approached for interview, unless they were too unwell (e.g. experiencing severe physical pain and unable to move, on a drip, vomiting and diarrhoea). The ward manager would indicate to the researcher which women were too unwell to be approached. All women on the postnatal ward were approached for interview unless they were too unwell or had delivered a stillbirth. Women who had a caesarean section were approached three to five days after surgery.

Consent to take part in the survey was obtained sequentially. Women were initially informed by the researcher (LB) that the research was to assess women's health during pregnancy. Women who agreed to participate were taken to a private room away from the wards where they were given a full explanation of the study. Women were informed that all information was confidential, unless they indicated that there was a risk of harm to an existing child. Participants were offered a £10 gift voucher in recognition of their assistance.

All women who participated in the research were offered contact cards with information about local organisations that offer support to women experiencing domestic violence. Additional time was given to women who disclosed domestic violence and wished to discuss this further. Women were also asked whether they wished any disclosure of domestic violence or abuse to be communicated to a health professional.

A semi-structured interview included questions about demographics and psychological health. Civil status was coded as a dichotomous variable: ‘married or cohabiting’ and ‘single, separated, or in a non-cohabiting relationship’. Ethnicity was collapsed into three categories due to the small numbers in some ethnic groups: ‘Black’, ‘White’ and ‘other’. Employment was coded as ‘in paid employment/on maternity leave’ or ‘unemployed/in receipt of sickness benefit’. Socio-economic classification was derived using the Standard Occupational Classification, Volumes 1 and 214,15 using the categories: ‘managerial and professional occupations’, ‘intermediate occupations’ and ‘routine and manual occupations (including the long term unemployed)’. Experiences of domestic violence were assessed using a variation of the Abuse Assessment Screen.16 Women who reported physical or sexual violence by a current or former partner/husband or family member were coded as positive for a history of domestic violence.

With regard to psychological health, women were asked whether or not they had ever consulted their GP complaining of ‘nerves, anxiety, sleeping problems, or feeling sad’, if they had ever received a psychiatric diagnosis from a health professional, and treatment they were offered. Dichotomous ‘yes/no’ questions were used to elicit information about self-harming behaviour, suicidal thoughts and suicide attempts.

Women were additionally asked to complete two self-report psychological measures. The Edinburgh Postnatal Depression Scale (EPDS) is a 10-item scale to assess depression, which was originally developed for postnatal use,17 but is now validated for use during pregnancy.18 All women were asked to rate on a four-point scale how they felt in the preceding week in response to the 10 items. A cutoff score of 14/15 was used antenatally, as recommended by Murray and Cox18 because of the high levels of dysphoria in pregnancy, and 12/13 postnatally.

Lifetime exposure to traumatic events and posttraumatic stress disorder were assessed using the Posttraumatic Diagnostic Scale (PTDS). The PTDS is a self-report scale that measures lifetime exposure to traumatic events, as well as a providing a diagnosis of posttraumatic stress disorder and posttraumatic stress symptom severity in the past month. It also asks about the duration of disturbance and resulting impairment. The instrument has been validated against the SCID (Structured Clinical Interview for Posttraumatic Stress Disorder; DSM-IV) and demonstrates high internal consistency and test–retest reliability, and good sensitivity and specificity.19 A cutoff score of 17 is taken to indicate clinical ‘caseness’ for a diagnosis of posttraumatic stress disorder.

Prevalence rates are presented with appropriate 95% confidence intervals. The relationship between lifetime trauma and current depression and posttraumatic stress disorder was examined both descriptively and inferentially. The χ2 test was used to examine the association between outcome variables measured at the nominal level. Adjusted χ2 tests (Yates' continuity correction) was used for all 2 × 2 tables, except where there were expected counts less than 5, in which case the Fisher–Irwin exact test was used. Pearson's χ2 tests was used for all 2 ×n (n > 2) tables if >25% of expected counts were greater than 5, otherwise the Fisher–Irwin exact test was used. The t test was used to compare means between the groups for data measured at the ratio level. Although some of these variables appear to have a skewed distribution and therefore suggest the use of non-parametric alternatives, t tests and analysis of variance were felt to still be robust due to the large sample size. Multiple logistic regression was used with dichotomous dependent variables to generate odds ratios while controlling for confounders. Confounders were included in the regression model if they were significantly associated with the dependent variable univariately at P≤ 0.1. This level was set on the basis of statistical advice before the data analysis. Analysis of variance was used to compare the mean EPDS scores among different groups. The Statistical Package for the Social Sciences (SPSS version 10.0 for Windows) was used to store and analyse the data.20


Between July 2001 and April 2002, 1144 women on the postnatal and antenatal wards and Day Assessment Unit were invited to participate in the study. Of the women approached, 200 (17.4%) women agreed to participate, 106 women from the postnatal ward and 94 women from the antenatal ward and Day Assessment Unit (Table 1).

Table 1.  Socio-demographic characteristics of antenatal and postnatal women. Values are presented as n (%).
Socio-demographic characteristicsAntenatalPostnatalTotal or mean ageχ2 or t test (df)P
Civil status
Married/cohabiting72 (76.6)88 (83.0)160 (80.0)0.9 (1)0.33
Single, separated, in a relationship, not living together22 (23.4)18 (17.0)40 (20.0)  
White57 (60.6)55 (51.9)112 (56.0)3.0 (2)0.21
Black31 (33.0)37 (34.9)68 (34.0)  
Other6 (6.4)14 (13.2)20 (10.0)  
Employment status
Paid employment/maternity leave58 (61.7)69 (65.1)127 (63.5)0.1 (1)0.72
Unemployment/disability benefit36 (38.3)37 (34.9)73 (36.5)  
Socio-economic classification
Managerial and professional26 (27.7)35 (33.0)61 (30.5)1.6 (2)0.44
Intermediate occupations17 (18.1)23 (21.7)40 (20.0)  
Routine and manual occupations/unemployed51 (54.3)48 (45.3)99 (49.5)  
Mean age [SD]29.5 [5.9]31.5 [6.0]30.6 [5.8]2.4 (198)0.02

Of the 200 women surveyed, 47 (23.5%; 95% CI = 18.2–29.8%) reported a history of domestic violence on the Abuse Assessment Screen. There was no significant difference in the reporting of domestic violence between postnatal and antenatal women (28/94; 29.8%vs 19/106, 17.9%; χ2= 3.3, df= 1, P= 0.07). The perpetrator of the abuse was a current partner or husband (5, 10.6%; 95% CI = 4.6–24.6%), an ex-partner or husband (35, 74.5%; 95% CI = 60.5–84.7%) and a family member or relative (7, 14.9%; 95% CI = 7.1–26.7%). One hundred and twenty-one (60.5%; 95% CI = 53.6–67.0%) women had been exposed to at least one traumatic event on the PTDS and 78 (64.5%; 95% CI = 55.6–72.4%) of these women had experienced multiple traumatic events.Table 2 presents the lifetime exposure to traumatic events amongst the 200 women.Table 3 presents the associations between domestic violence and socio-demographic factors.

Table 2.  Socio-demographic characteristics by history of domestic violence (Abuse Assessment Screen). Values are presented as n (%).
Socio-demographic characteristicsHistory of domestic violence (n= 47)No history of domestic violence (n= 153)χ2 or t test (df)P
  • *

    If blank, then Fisher–Irwin exact test is used.

Civil status
Married/cohabiting29 (38.3)22 (14.4)11.4 (1)0.001
Single, separated, in a relationship, not living together18 (61.7)131 (85.6)  
White23 (48.9)89 (58.2)2.0 (2)0.4
Black20 (42.6)48 (31.4)  
Other4 (8.5)16 (10.5)  
Employment status
Paid employment/ maternity leave25 (53.2)102 (66.7)*0.1
Unemployment/ disability benefit22 (46.8)51 (33.3)  
Socio-economic classification
Managerial and professional12 (25.5)49 (32.0)0.7 (2)0.7
Intermediate occupations10 (21.3)30 (19.6)  
Routine and manual occupations/unemployed25 (53.2)74 (48.8)  
Mean age [SD]30.6 [5.9]30.3 [5.8]0.3 (198)0.7
Table 3.  Lifetime exposure to traumatic events of 200 women. Values are presented as n (%).
Eventn (%)
Serious incident, fire, or explosion (e.g. an industrial, farm, car, plane or boating accident)38 (19.0)
Natural disaster (e.g. tornado, hurricane, flood or major disaster)19 (9.5)
Physical assault by a family member or someone you know (e.g. being mugged, physically attacked, shot, stabbed or held at gunpoint)68 (34.0)
Physical assault by a stranger (e.g. being mugged, physically attacked, shot, stabbed or held at gunpoint)34 (17.0)
Sexual assault by a family member or someone you know (e.g. rape or attempted rape)22 (11.1)
Sexual assault by a stranger (e.g. rape or attempted rape)7 (3.5)
Military combat or a war zone4 (2.0)
Sexual contact when you were younger than 18 with someone who was 5 or more years older than you (e.g. contact with genitals, breasts)20 (10.1)
Imprisonment (e.g. prison inmate, prisoner, or war hostage)15 (7.5)
Torture8 (4.0)
Life-threatening illness40 (20.0)
Other traumatic event25 (12.5)

Women who reported a history of domestic violence were significantly more likely to have experienced at least one other traumatic event compared with women with no history of domestic violence (33/47, 70.2%vs 74/153, 48.4%; χ2= 6.0, df= 1, P= 0.01).Table 4 presents the associations between domestic violence and traumatic events. Only childhood sexual abuse and physical or sexual assault by an unknown assailant was significantly associated with a history of domestic. Accidental or non-assaultive traumas were no more likely to be reported by victims of domestic violence than by women with no history of domestic violence. Unwanted sexual experiences as a child were significantly associated with physical or sexual assault by any perpetrator as an adult (16/20, 80.0%vs 69/179, 38.5%; χ2= 11.0, df= 1, P= 0.001) and also with exposure to any adult trauma, unrelated to interpersonal violence (14/20, 70.0%vs 73/179, 40.8%; χ2= 5.1, df= 1, P= 0.02).

Table 4.  Associations between domestic violence and other traumatic events. Values are presented as n (%).
Traumatic eventDomestic violence (n= 47)No domestic violence (n= 153)χ2 (df)*P
  • *

    If blank, then Fisher–Irwin exact test is used.

Serious incident, fire or explosion (e.g. an industrial, farm, car, plane or boating accident)11 (23.4)27 (17.6)0.4 (1)0.5
Natural disaster (e.g. tornado, hurricane, flood or major disaster)8 (17.0)11 (7.2)3.0 (1)0.08
Physical or sexual assault by a stranger (e.g. being mugged, physically attacked, shot, stabbed or held at gunpoint)14 (29.8)24 (15.7)3.8 (1)0.05
Military combat or a war zone0 (0.0)4 (2.6) 0.6
Sexual contact when you were younger than 18 with someone who was (≥5 years older (e.g. contact with genitals, breasts)10 (21.3)10 (6.6)7.0 (1)0.008
Imprisonment (e.g. prison inmate, prisoner or war hostage)6 (12.8)9 (5.9)*0.1
Torture6 (12.8)2 (1.3) 0.002
Life-threatening illness11 (23.4)29 (19.0)0.2 (1)0.6
Other traumatic event7 (14.9)18 (11.8)0.09 (1)0.8

In a logistic regression, both domestic violence (OR = 2.7, 95% CI = 1.3–5.8, P= 0.008) and lifetime experience of trauma, excluding domestic violence, emerged as significant predictors of women having ‘ever visited a GP with nerves, anxiety, problems sleeping or feeling sad’ (OR = 3.5, 95% CI = 1.6–7.7, P= 0.002).Table 5 presents the univariate associations between psychological health and history of trauma and psychological health and domestic violence.

Table 5.  Associations between psychological health, lifetime trauma and domestic violence. Values are presented as n (%). χ2 test was used to obtain P values.
Psychological health outcomesTrauma (n= 121)No trauma (n= 79)χ2 (df)PDomestic violence (n= 47)No domestic violence (n= 153)χ2 (df)*P
  • *

    If blank, then Fisher–Irwin exact test is used.

Suicidal thoughts43 (35.5)3 (3.8) 0.00115 (31.9)31 (20.3)2.1 (1)0.1
Suicide attempts12 (9.9)1 (1.3) 0.035 (10.6)8 (5.2) 0.2
Self-harm6 (5.0)0 (0.0) 0.082 (4.3)4 (2.6) 0.6
Diagnosed with depression by a health professional30 (24.8)5 (6.3) 0.0029 (19.1)26 (17.0)0.01 (1)0.9
Consulted a GP with ‘nerves, anxiety, problems sleeping or feeling sad’38 (31.4)7 (8.9) 0.00119 (40.4)26 (17.0)10.0 (1)0.002

There was no significant association between clinical ‘caseness’ on the EPDS and having a history of trauma, multiple traumatic events, domestic violence, trauma history involving adult or childhood physical or sexual violence, or being admitted for obstetric complications. However, there were significant differences within these groups regarding mean EPDS scores (see Table 6).

Table 6.  Comparison of mean EPDS scores. T test was used to obtain P values.
Groups comparedMeanSDDifference in means95% CIP
  • *

    Interpersonal violence defined as childhood and/or adult sexual or physical violence or torture.

Any trauma history vs9.–3.30.02
No trauma history7.44.7   
Multiple traumatic events vs9.–4.30.04
Single traumatic event7.75.2   
History of domestic violence vs10.–4.30.004
No domestic violence7.85.1   
Trauma related to interpersonal violence vs9.–4.2<0.001
Traumas unrelated to interpersonal violence*7.25.6   
Antenatal group vs9.–4.10.001

Analysis of variance was used to compare mean EPDS scores between the following five dichotomous groups: history of domestic violence, history of trauma, history of trauma involving adult/childhood physical or sexual violence, ward (antenatal vs postnatal status), and married or cohabiting vs single, separated or in a non-cohabiting relationship. Only being single, separated or in a non-cohabiting relationship emerged as significant in the model (F= 4.9, df= 1, P= 0.03). Because women with a history of abuse were significantly more likely to be single, separated or in a non-cohabiting relationship, this procedure was repeated, removing civil status from the model. A history of trauma involving adult/childhood physical or sexual assault (F= 3.9, df= 1, P= 0.05) and being on the antenatal ward (F= 8.1, df= 1, P= 0.005) were associated with significantly higher EPDS scores using this model.

The mean PTDS score for the 121 women who had experienced any trauma was 4.3 (SD = 8.0, range 0–45). Thirteen (10.7%) of the 121 women scored 17 or more, corresponding to a clinical diagnosis of posttraumatic stress disorder. Women exposed to multiple traumatic events scored significantly higher on the PTDS compared with women exposed to a single traumatic event (6.5, SD = 10.5 vs 0.3, df= 1.1; diff = 6.2, 95% CI = 3.9–8.7). Women exposed to a traumatic event that had occurred between one month and three years previously scored significantly higher on the PTDS compared with women exposed to a traumatic event that occurred more than three years ago (8.1, SD = 11.4 vs 2.3, df= 6.6; diff = 5.8, 95% CI = 2.0–9.6). The women with a trauma history that involved physical or sexual violence as a child and/or adult scored significantly higher on the PTDS (5.1, sd = 9.6 vs 1.8, SD = 6.2; diff = 3.3, 95% CI = 0.3–6.3) compared with women who reported other forms of trauma. Clinical caseness on the PTDS was significantly associated with caseness for depression as measured by the EPDS (χ2= 5.9, df= 1, P= 0.01). Seven (25.0%) of the 28 women with caseness on the EPDS also had a score of 17 or more on the PTDS, while only 6 (6.5%) of the 93 women who scored below clinical caseness on the EPDS met clinical caseness on the PTDS.


The response rate of 17% is rather low and reflects the practical difficulty associated with creating the time and private space to interview newly delivered mothers on busy postnatal wards has been noted by other researchers.21 We do not know whether there was any systematic difference between the women who declined to take part when first approached on the ward and those who agreed, as we did not collect demographic or obstetric data on women who were not involved in the study. However, we think that the relatively low rate of compliance is unlikely to have biased the results, as women were only told what the study was about, after they had accompanied the researcher to a private room and had agreed to participate in the study. The main reasons women gave for not doing the survey were that they were too tired, were expecting or entertaining visitors, were preoccupied with the baby (postnatal ward), worried about the baby's health or simply getting ready to go home. Women were often anxious that the ward clerks and midwives complete all the relevant paper work so they could be discharged quickly and did not want to be bothered by research. There is no evidence that women were reluctant to be questioned about domestic violence, as none of the women who initially agreed to be questioned by the researcher refused to give their consent or withdrew from the study once the specific nature of the enquiry was described to them.

The psychological measures used in this study only provide information about current levels of depressive and posttraumatic stress symptomatology. We did not examine lifetime diagnoses of depression or posttraumatic stress disorder, or their relation to trauma. In addition, the antenatal group only included women who had been seen or admitted as a result of obstetric complications and there fore, the findings in this group may not be applicable to women with an uncomplicated pregnancy. It is possible that some of the current depressive symptomatology in the antenatal group reflected the experience of being admitted to hospital. However, there was no significant difference in rates of clinical caseness for depression in the antenatal and postnatal groups, and the levels of depressive symptoms reported in this study are comparable to other studies of depression in ‘uncomplicated’ antenatal populations.22

Traumatic life events were reported by a substantial number of women and were comparable to that described in studies of women outside maternity services.4,5,23 Around one in 10 women reported contact sexual abuse as a child and one in four women had experienced domestic violence at some point in their lives, which is consistent with a number of UK24 and Irish surveys,25 but somewhat lower than the 41% lifetime rate of domestic violence recently reported by Richardson et al.26 in women attending GP surgeries in East London.

This study identified a strong association between a trauma history and current depressive and posttraumatic stress symptoms. The majority of women in this study had experienced more than one traumatic event in their lives and these women had higher levels of depressive and posttraumatic stress symptoms than women who had only suffered a single trauma. Thus, it would appear that the psychiatric effects of these traumas are cumulative.27 Higher levels of depressive and posttraumatic stress symptoms were also found in victims of interpersonal violence and abuse than other forms of trauma. Around one in 10 women who had experienced any lifetime trauma met the criteria for a current diagnosis of posttraumatic stress disorder, a rate which is similar to that reported in other studies.28

Domestic violence affects around one in three women over a lifetime29,30 and one in nine on an annual basis.31 A recent study of women attending General Practice surgeries in East London10 reported lifetime rates of 41% for domestic violence, 4% for rape outside an intimate relationship, and 21% of women report ‘other’ traumatic experiences. Women who experience domestic violence are at increased risk of developing psychiatric illness including depression, anxiety, substance misuse and posttraumatic stress disorder.32,33 In this study, childhood sexual abuse and single or cohabiting civil status was also associated with depressive symptoms. It is recognised that victims of physical and sexual abuse are more likely to experience problems in terms of their intimate and marital relationships34 and to be single, separated or divorced24 which means that, in practice, they often present with depressive symptoms in clinical settings.

The psychological impact of being single, separated or in a non-cohabiting relationship may be different for women who are pregnant or who have recently given birth, than for non-pregnant women, so the results of this study cannot necessarily be generalised to women outside maternity services. Pregnancy and the immediate period following birth is a time when partner support and being in a confiding trusting relationship may be particularly important for psychological health. Women may feel more physically vulnerable, and emotionally dependent on their partners during pregnancy and after giving birth. In addition, the implications of domestic violence are perhaps more profound during pregnancy and the immediate postpartum period because of the threat such violence poses to the safety of both mother and baby.

Women with a history of interpersonal victimisation may experience pregnancy and childbirth as particularly threatening to their physical integrity. Obstetric procedures which involve submitting to repeated intimate and intrusive physical examination, often by male doctors, may intensify feelings of helplessness and entrapment, which are recognised as risk factors in the development of depression35 and reactivation of posttraumatic stress symptoms.36 Although posttraumatic stress disorder has been identified as a potential response to childbirth and antenatal complications,37,38 none of the women in this study selected birth or any aspect of their current pregnancy as the index traumatic event.

Coid et al.10 have suggested that the association between childhood sexual abuse and adult re-victimisation is mediated by socio-economic factors and increased alcohol use by victims. Unlike Coid's research, this study found no significant association between domestic violence and socio-economic disadvantage, or increased alcohol consumption. The results of this study demonstrate that the vulnerability of victims of childhood sexual abuse to re-traumatisation extends beyond interpersonal traumas. It may be that childhood sexual abuse leads to alterations in the victim's capacity for self-protection and self-efficacy, and self-esteem, as well as the way they perceive, judge or respond to environmental threat.24,39 Some of the psychiatric effects of childhood sexual abuse, including personality disorders and posttraumatic psychiatric illness, would also increase the risk of trauma exposure.40

Domestic violence rarely exists in isolation and probably represents a marker for a range of childhood and adult abuse experiences. Our findings support the conclusions of Mullen et al.,41 who argued that the psychiatric and social effects of childhood abuse represent a ‘matrix of developmental disadvantage’ which includes sexual abuse, relationship dysfunction and repeat victimisation. It is important therefore that any apparent association between domestic violence and psychiatric illness is not attributed to the effects of domestic violence in isolation, but understood in this context.


This study was funded by the Economic and Social Research Council. The authors would like to thank the women who participated in the survey.

Accepted 26 April 2004