Case definition

The study examined experience of domestic violence, which should be contrasted with perpetration. Domestic violence was defined by a ‘yes’ response to a number of questions that described domestic violence, including physical, sexual and emotional violence by the participants' current and former partner. The questionnaire applied the Home Office case definition of domestic violence, being ‘any violence between current and former partners in an intimate relationship, wherever the violence occurs. The violence may include physical, sexual, emotional and financial abuse’ (www.justice.gov.uk/about/domesticviolence.htm). Women were also asked about the acceptability of being questioned by healthcare workers about domestic violence by a series of questions, for example: ‘In general, would you mind if your psychiatrist asked you whether you were being threatened, hit or hurt by your partner or a previous partner?’ The self-administered questionnaire was piloted and modified, and thereafter completed in the presence of the researchers.

Data were entered, checked for accuracy and analysed using SPSS version 14 for Windows. Univariate analyses were conducted using the χ²-test for frequencies or Fisher's exact test where appropriate. Logistic regression analysis was used to identify risk factors relating to domestic violence. Sample size calculations showed that 64 participants were needed to have 90% power to show a 15% difference in a range of demographic variables and to be significant at the 0.05 level between women who had experienced physical violence.

Participants' ethnic groups were based on categories defined by the Office for National Statistics. ^{Reference Dobbs, Green and Zealey8} We restricted the study to female patients because of the high prevalence of domestic violence among women in the general population and the gender of the researchers, as suggested by the WHO ethics guidelines for research into domestic violence. ⁹

The questions looked at: demographic information including past or present relationship; psychiatric diagnosis; self-harm; history of childhood abuse; different categories of domestic violence; women's attitudes to being questioned by psychiatrist or key worker about abuse; and any previous disclosure. Participants were invited to record their comments about the study. Data were collected on diagnoses and any disclosure of domestic violence documented in the case notes.

Prevalence

Experiences of specific categories of domestic violence are shown in Table 1. We make reference predominantly to lifetime data as only a small number of women had a current partner in the past 12 months. However, Table 1 includes some data on violence in the past 12 months, within the limits of a small data-set. In total, 60% of women (n = 42) had experienced physical violence from their current or previous partner, 81% (n = 57/70) had experienced some form of controlling behaviour and 54% (n = 38) had been threatened by their current or previous partner. A smaller proportion (52%, n = 37) responded positively to the question ‘Do you think you have ever experienced domestic violence?’ Domestic violence was not uncommon during pregnancy and was reported by 14 of the study population (n = 70, 27%), of whom a quarter ascribed worsening violence to the pregnancy and 12% (n = 3/25) attributed miscarriage to the violence.

Table 1 Prevalence of domestic violence in the study sample with 95% CI (n = 70)^a

Type of violence	n (%)	95% CI
Do you think you have ever experienced domestic violence?	37 (52)	40.6–64.9
Physical violence by partner	42 (60)	47.6–71.5
Grabbed or shoved you	31 (43)	32.4–56.7
Punched you on body/arms/legs	25 (36)	24.6–48.1
Punched you in the face	19 (27)	17.2–39.1
Forced you to have sex	24 (33)	23–46.6
Physically violent to you in other way	20 (27)	18.4–40.6
Kicked you on the floor	16 (24)	13.7–34.4
Choked or held hand over your mouth	19 (26)	17.2–39.1
Used weapon or object to hurt you	13 (19)	10.3–29.7
Tried to strangle, burn or drown you	10 (13)	24.7–70.7
Threatening behaviour by partner	38 (54)	41.9–66.3
Punched, kicked or threw things	31 (44)	32.4–56.7
Threatened you with fist, hand or foot	29 (40)	29.8–53.8
Threatened you with object or weapon	20 (29)	18.4–40.6
Threatened to kill you	17 (24)	14.8–36
Controlling behaviour by partner	57 (81)	70.3–89.7
Shouted, screamed or swore at you	48 (69)	56.4–79.1
Criticised you	52 (74)	62.4–84
Checked up on your movements	31 (44)	33.7–58.1
Restricted your social life	37 (53)	40.6–64.9
Tried to control you in any other way not physical violence	34 (49)	37.8–62.2
Kept you short of money	30 (43)	29.8–53.8
Locked you in the house	10 (15)	24.7–70.7
Physical violence in the past 12 months	11 (15)	26.4–81.1
Have you ever felt afraid of your partner?	41 (59)	46.2–70.2
Domestic violence during pregnancy	14 (27)	11.4–31.3
True rate of recording of domestic violence in case notes	18 (26)	15.8–37.1

Disclosure and recording in the clinical notes

Of 65 women who had experienced domestic violence, 40% (n = 26) reported physical injuries as a direct result but only half sought medical attention. Regarding disclosure, of 39 women who answered the question, 51% reported disclosure of domestic violence to the psychiatrist and 29% to their general practitioner (GP). The majority of participants had never been questioned by a psychiatrist or other health professional about experiences of domestic violence, with only 22% (n = 16) having received questions from a psychiatrist or key worker about threatened or actual physical violence and 13% (n = 9) about sexual violence by their partner. Addressing attitudes and acceptability of such questioning, the vast majority (n = 58, 82%) regarded questions from a psychiatrist or key worker as entirely acceptable, with 10% objecting and the same proportion ‘unsure’. There were no demographic or clinical differences between those objecting and those not, which may represent a type II error. Gender of psychiatrist was considered relevant by most participants, with 53% (n = 38) expressing a preference for a woman and 39% (n = 28) unconcerned by gender. Domestic violence as a precipitant of hospital admission was also examined by medical note review, but no record of this was found. In only 18 of the women's case notes (25%) was there any record of domestic violence.

Variables associated with domestic violence

There was no statistically significant association between domestic violence and age, employment, ethnic group or being born in the UK (online Table DS1). There were significant associations with having experienced actual physical violence (as opposed to threatening or controlling behaviour) (n = 42, 60%), having taken an overdose (n = 22, 54%) and having children (n = 32, 78%). Cohabiting with a partner was protective (n = 7, 19%). Admission to psychiatric hospital or being detained under the Mental Health Act did not put women at more risk of domestic violence.

Sexual violence

About half of those women who had suffered violence (n = 23, 55%) had also experienced sexual violence by their partner. In addition, 18 women (25%) had experienced sexual abuse before the age of 16 and 25 (35%) since the age of 16 years (online Table DS1).

Interrelationships between variables

Logistic regression analyses were used to identify demographic variables that were significantly related to domestic violence. For the purpose of this analysis, we included any woman who had ever experienced any type of physical violence. We defined current domestic violence as physical violence experienced during the past 12 months. We constructed a logistic regression model to assess the influence of all those variables found to be significant in influencing outcome (physical violence): presence of children, a history of overdose, and experience of sexual abuse before and after the age of 16 years. Variables were entered in a forward stepwise manner; they were considered as significant at the 5% level. The presence of children, a history of overdose, and experience of sexual abuse after the age of 16 years at presentation were the only variables found to have a significant independent effect on experiencing physical violence (Table 2).

Table 2 Logistic regression analysis of characteristics associated with domestic violence

	P	Odds ratio (95% CI)
Children	0.05	3.34 (0.97–11.49)
Overdose	0.02	4.32 (1.24–14.93)
Unwanted sexual experience since age 16 years	0.002	6.76 (2.00–22.73)

Disclosure

Despite high prevalence, domestic violence was only recorded in clinical notes in 26% of cases, although the majority of women found questions regarding violence acceptable.

Our study replicates findings in primary care ^{Reference Richardson, Coid, Petruckevitch, Chung, Moorey and Feder2} in terms of acceptability of routine enquiry for domestic violence: most women were in favour, with a minority objecting. Health professionals show reticence towards enquiring about violence; they do not feel trained or equipped for effective interventions. ^{Reference Ramsay, Richardson, Carter, Davidson and Feder10} Concerns have been expressed that routine enquiry may also endanger the women, such that it has not been recommended except in high-risk groups such as pregnant women.

This study suggests that women with mental disorders should also be included in high-risk groups, and psychiatric health workers should be equipped to enquire about violence. In a systematic review, Ramsay et al ^{Reference Ramsay, Richardson, Carter, Davidson and Feder10} suggest that routine enquiry without evidence of effective interventions should not be recommended. On the contrary, this study supports routine enquiry in psychiatric services. In our sample, 69% of the women had been admitted to a psychiatric hospital and 30% had been detained under the Mental Health Act, indicating the severity of their illnesses and probably their increased vulnerability to become victims of violence. These findings alone should justify, from a human rights perspective and on ethical grounds, enquiries about domestic violence. In addition, the Royal College of Psychiatrists' policy statement on domestic violence ¹¹ recommends that enquiry on domestic violence should be part of the routine clinical assessment. The Victims of Violence and Abuse Prevention Programme has established a pilot to introduce routine clinical enquiry about violence and abuse as a component of the care programme approach. ^{Reference Itzin, Bailey and Bentovim12}

Risk factors

Univariate analysis showed that those abused by their partner were more likely to have attempted suicide by taking overdoses, to use street drugs, to have had children and to have suffered sexual abuse before and since the age of 16 years. The significant association between physical abuse and a history of overdose confirms data from other studies of high levels of depression as a consequence of domestic violence. ^{Reference Straus and Gelles13,Reference Coid, Petruckevitch, Feder, Chung, Richardson and Moorley14} The particular risk to mothers highlights the need to consider issues of safeguarding children in such cases. The literature has identified that women under 40 years of age who are single, separated or divorced are at higher risk. ^{Reference Dearwater, Coben, Campbell, Nah, Glass and McLaughlin15} We did not find any association between violence and age. This might be due to the fact that the study population was older and that only a small number were in a relationship (n = 27, 38%).

The history of sexual abuse after the age of 16 replicates the findings of the study by Coid et al. ^{Reference Coid, Petruckevitch, Feder, Chung, Richardson and Moorley14} We cannot assert any causal link in view of study design, small sample and lifetime prevalence as outcome measure. However, a disproportionately high number of women suffering sexual abuse is worthy of note: 18 before the age of 16 and 25 after.

A smaller proportion (52%, n = 37) responded positively to the question ‘Do you think you have ever experienced domestic violence?’, which replicates findings in a primary care setting. ^{Reference Richardson, Coid, Petruckevitch, Chung, Moorey and Feder2} Women who have suffered violence at times do not recognise it as such, or may think they ‘deserved’ it. ^{Reference Richardson, Coid, Petruckevitch, Chung, Moorey and Feder2}

Physical injuries

Physical injuries were reported by 42% of the women, but only half sought medical attention. These are very high rates, explicable in a psychiatric population by the impact of physical injuries on mental health, and by the vulnerability of women with severe mental disorders. However, because of the small numbers we cannot attribute causality.

Study limitations

We cannot be confident that our results generalise to other psychiatric populations, although the findings replicate those of other studies and we achieved a high response rate for work in this field. We have not examined prevalence among male psychiatric patients, and comparable studies are needed. The study demonstrates associations, but it was not constructed to test hypotheses concerning causal relations. The study was not controlled and it is possible that higher rates of violence reflect the London population from which the sample was drawn.

This study demonstrates that domestic violence in female psychiatric patients is much more common than in other patient groups. Most female psychiatric patients have experienced domestic violence but almost half have not disclosed it. Disclosure to a psychiatrist is rare, but acceptable to most patients. We suggest that routine enquiry about domestic violence should be considered in women who experience serious mental illness, and women should receive help with domestic violence as part of their care. The nature of that help would be determined by circumstances, but should include social and legal advice, as well as emotional support focused on safety planning.