Gendered experiences of physical restraint on locked wards for women

Abstract Physical restraint is used in inpatient services for people with intellectual disabilities as a way of holding a person to avoid injury. This article uses data from an ethnographic study in a locked unit in the north of England to explore women’s experiences of physical restraint using a feminist disability studies analysis. Data consists of field notes as well as interviews with 16 of the women who had experienced restraint, and 10 staff who worked with them. The women gave insights into the gendered phenomenon of restraint in light of their past experiences of violence. The authors argue that restraint is used with women to encourage passivity at times when more relational and therapeutic methods could be used. The article offers recommendations for alternative strategies that services can encourage.


Points of interest
• This article investigates the experiences of physical restraint of women with intellectual disabilities who lived in locked wards and their staff. • The women wanted more information about when restraint would happen and the reasons why they were restrained. • Many of the women reported that being restrained made them more angry and some said that it brought back bad memories of violence. • They recommended that staff talk to them more about why they are angry before restraint is needed. • Good relationships between staff and service users can reduce the need for restraint.

Introduction
This article describes a theme that arose from a research project with intellectually disabled women detained in a National Health Service (NHS) inpatient forensic unit in the north of England. The project used ethnography to explore the daily life of staff and detained women on three of the wards at an NHS intellectual disability secure unit in the north of England ('Unit C'). Physical intervention or restraint is used on such units when people detained in them become 'unmanageably aggressive' . It is used to restrict someone's freedom of movement when there is the threat of injury to others or damage to property. The Mental Health Act code of practice defines the practice of physical restraint as 'any direct physical contact where the intention is to prevent, restrict, or subdue movement of the body (or part of the body) of another person' (Department of Health 2007, 295). The document advises that restrictive interventions: should never be employed to deliberately punish or humiliate, and staff should not cause deliberate pain to a person in an attempt to force compliance with their instructions except in the most exceptional circumstances to mitigate an immediate risk to life. (Department of Health 2007, 294) The document also specifies that: where possible, the patient should be asked to stop the behaviour … an explanation should be given of the consequences of refusing the request from staff to stop the behaviour. The explanation should be provided calmly and every attempt should be made to avoid the explanation being perceived by the patient as a threat. (Department of Health 2007, 294) In this way, UK policy makes it clear that the circumstances for using restraint should be reactive rather than proactive, and used as a last resort, thereby avoiding the person experiencing restraint as punishment. This is a consistent refrain of guidance concerning restraint in services for people with intellectual disabilities, accompanied by recurring messages that rates of physical restraint are too high and the types of physical restraint used are unnecessarily harmful (Department of Health 2002, 2014. The experiences of people with intellectual disabilities suggest that practice differs from policy guidance. Griffith, Hutchinson, and Hastings (2013) produced a synthesis of studies investigating service responses to 'challenging behaviour' . They found that people with intellectual disabilities felt powerless due to the negative atmosphere on wards and they experienced a loss of autonomy and violence or indifference from others, which engendered a cycle of frustration and further challenging behaviour. Griffith, Hutchinson, and Hastings concluded that although people understood the reasons why restrictive interventions were used, they nevertheless experienced them as a form of punishment. Furthermore Sequeira and Halstead (2001) found that the use of physical intervention could cause pain and be construed as punishment, and participants in Fish and Culshaw's (2005) study described feeling retraumatised by the episode of restraint; in that being overpowered by another person reminded them of violence and abuse in their pasts. Such power differentials arguably reflect broader cultural dynamics towards disabled people in general, who experience widespread discrimination, marginalisation, social exclusion and disempowerment.

Gender and restraint in institutions
Research investigating the use of restraint (broadly defined) in services for people with intellectual disabilities has rarely considered restraint in the context of gender. Quantitative studies focusing on restraints used with people with intellectual disabilities in community settings have reported high proportions of people being subject to chemical, physical and mechanical restraint, and a minority of people being subject to frequent and multiple restraints (for example, Emerson et al. 2000;Lundström et al. 2011;Merineau-Cote and Morin 2013;Webber, McVilly, and Chan 2011). However, these studies have not routinely included gender analyses and have tended to conclude that gender differences are relatively minor. Studies of restraint, seclusion and use of PRN (pro re nata or 'as needed') psychoactive medication used with people with intellectual disabilities in inpatient services have also rarely included gender analyses (Scheirs et al. 2012;Sturmey 2009;Tilli and Spreat 2009). A recent systematic review of the research literature concerning how people with intellectual disabilities and staff members experienced restraint also did not consider gender (Heyvaert et al. 2014). Finally, studies reporting programmes designed to reduce the use of restraint in services for people with intellectual disabilities have not considered gender in their design or reporting (Williams and Grossett 2011).
Two quantitative studies have included gender analyses in their research concerning restraint in inpatient services. Firstly, a survey collecting information about 268 people with autism and/or intellectual disabilities in specialist inpatient services for people with intellectual disabilities in England (McGill, Murphy, and Kelly-Pike 2009) reported that women were equally likely to have been subject to seclusion (27% of women compared with 26% of men) and less likely to have been subject to physical restraint (74% of women compared with 88% of men). They were more likely than men to have been subject to 'protective devices' and to have had psychoactive medication administered 'PRN' .
However, a more recent study involving a census of inpatient services for people with intellectual disabilities in England in 2015 (NHS Digital 2015) revealed stark differences in the experiences of men and women. Women with intellectual disabilities constituted a quarter of the 3000 people in inpatient services. Women were more likely to have experienced at least one 'adverse incident' (self-harm, accident or physical assault against them) in the three months before the census (62% of women compared with 41% of men), and were also more likely to have experienced at least one 'restrictive measure' (hands-on restraint, seclusion) in the last three months (47% of women compared with 32% of men). In addition, women were more likely to have been administered tranquilising medication and antipsychotic medication in the 28 days before the census.
This census further showed that adverse incidents and restrictive measures were particularly prevalent in women-only wards. For example, 67% of women in single-sex wards experienced an adverse incident in the last three months, compared with 54% of women in mixed wards, 41% of men in male-only wards and 39% of men in mixed wards. Similarly, 54% of women in single-sex wards experienced a restrictive measure in the last three months, compared with 34% of women in mixed wards, 33% of men in male-only wards and 29% of men in mixed wards.
These studies reveal the extremely widespread use of restrictive practices to control behaviour within inpatient services. While they do make important points, they do not go as far as analysing the gendered effects of these restrictive practices. Although some qualitative projects explore gendered rationales for aggression (Clements, Clare, and Ezelle 1995;Edwards 1999;Hejtmanek 2010;Wilcox, Finlay, and Edmonds 2006), and some studies in mental health units have included women's experiences of restrictive practices, including feelings of retraumatisation (Gallop et al. 1999;Jennings 1994;Mohr, Petti, and Mohr 2003), no studies to date have focused solely on analysing experiences of physical restraint from a gender and disability perspective.
There are a number of studies that explore experiences of restraint in intellectual disability units (Fish and Culshaw 2005;Jones and Kroese 2007;Macdonald, McGill, and Deveau 2011;Mérineau-Côté and Morin 2014;Sequeira and Halstead 2001, 2002a, 2002b and some conclude that women experience restraint in particular ways, acknowledging the distinct potential for retraumatisation and powerlessness. This article extends the analysis to focus on women's embodied experiences in a context of a power structure that incorporates gender and disability. We will argue that the use of restraint is experienced as violence by disabled women, many of whom have experienced violence in their pasts.

Restraint policy
Unacceptably high levels of restraint (mechanical, physical, chemical and seclusion) in use in services for people with intellectual disabilities have been identified as an object of policy attention in several countries, often in response to media exposure of restraint practices (Romijn and Frederiks 2012). National policies typically have the aim of reducing restraint but not eliminating it, to reach a point 'where physical interventions are only used as a last resort' (Department of Health 2014, 9). Components of national policies may include legislation, government endorsement of policy aims, programmes to encourage better leadership within service provider organisations, various forms of staff training, improved regulation and some form of monitoring (Rickard, Chan, and Merriman 2013;Romijn and Frederiks 2012). Despite this array of policies and action plans, it is clear that restraint practices are somewhat resistant to policy direction, with some countries even reporting steadily increasing rates of reported statutory restraint after the introduction of such policies (Søndenaa, Dragsten, and Whittington 2015) and specialist inpatient services reporting little change in commonly used restraint practices after the introduction of restraint reduction policies (Hatton 2016;NHS Digital 2015).
In England and Wales, national 'guidance' (i.e. without statutory force) on reducing restraint in services for people with intellectual disabilities (and recently people in mental health services) has been issued ( . This extensive and detailed guidance is striking for its almost complete absence of attention to gender issues and its reliance on behavioural and biological frameworks to direct restraint practices. For example, the 46-page Department of Health guidance document contains a single reference to issues that may be particularly salient to women in inpatient services: Individual risk factors which suggest a person is at increased risk of physical and/or emotional trauma must be taken into account when applying restrictive interventions. For example, this would include recognising that for a person with a history of traumatic sexual/physical abuse, any physical contact may carry an additional risk of causing added emotional trauma. (Department of Health 2014, 26) Similarly, the 20-page Royal College of Nursing report contains one reference to gender: People in custody, which includes a large percentage of women, may have undiagnosed and unmet needs: anxiety disorders, possibly learning disabilities, as well as psychotic disorders resulting from alcohol and drug misuse. (Royal College of Nursing 2016, 7) This contrasts with a clear statement about gender in relation to restraint in a mental health context: Both women and men may be re-traumatised by restraint that parallels past physical or sexual abuse. The gender of staff involved in restraint may be relevant and make this more likely to occur, or be worse if it does. Staff expectations of, and responses to, behaviour may differ for men and women too. (National Survivor User Network & Mind 2015, 15) The lack of acknowledgement of the importance of gender in the use of physical restraint (and more broadly in discussions of various forms of trauma such as sexual assault) is concerning; with this article, we aim to provide evidence on the importance of gender as well as disability in experiences of restrictive practices.

Method
This article is drawn from ethnographic research with intellectually disabled women in locked wards. Ethnographic researchers are concerned with the meanings people attach to things in their lives (Taylor and Bogdan 1998;Goodley 2001).
By observing people in context during their everyday lives, researchers are able to develop concepts from the data rather than testing preconceived models or theories. Some researchers in the intellectual disability field have used observational methods in their research exploring the experiences of people in locked wards (Hubert and Hollins 2006;Johnson 1998;Owen, Hubert, and Hollins 2008). Other researchers have used ethnographic methods to enable people who are less articulate to participate in research (for example, Booth and Booth 1996). The fieldwork involved the researcher observing daily life on three of the wards where the women lived, for 120 hours over a period of nine months. Sixteen women detained on the wards and 10 staff members (two men and eight women staff ) subsequently agreed to be interviewed. Two of the wards were classified as low secure (wards are locked but women are able to access other areas of the unit) and one was part of the medium-secure unit (wards are locked and women must stay within the two-storey enclosure at all times).

Access, consent and ethics
The research was given ethical approval from the NHS Local Research Ethics Committee. One of the prerequisites of approval was that written consent had to be obtained from all of the people involved, to allow field notes to be written about any observation. The women who participated in the research were all labelled as having mild to moderate intellectual disabilities and were able to give consent. They were informed that the information they gave would be anonymous and that their name would not be written down anywhere. Pseudonyms were given during the transcription process and attempts were made to exclude any information that could identify participants (Ellis and Bochner 2000;Rogers and Ludhra 2012).
Ethical adherence followed the guidance of Perry (2004) who advises that ethical concerns should be paramount throughout the research. Case managers were consulted before requesting consent, and those women who were not able to give consent were not observed or recorded in any field notes. Separate consent forms were completed by staff and detained women, and for participation in the observation and interviews. Verbal checks were performed with the women to make sure they had understood the process of the research. Any participant who disclosed that they had been subject to abuse was referred for counselling if the issue had not already been dealt with.

Analysis
During the observational period, the field notes were transcribed and printed out and the printouts were cut with scissors into items of text. The research team spent many hours together arranging these items into themes. From these themes, the research team devised interview questions that arose directly from the observations. Twenty-six people agreed to be interviewed, 10 of them staff and 16 were women who were detained at the unit. The interview transcripts were anonymised and all of the data were uploaded to a qualitative software package, NVIVO, to aid with the management of such large amounts of data. The analysis method was borrowed from the phenomenological research tradition (Hycner 1985) and the steps suggested by Hycner were followed in order that the analysis was inductive, arising from the data. Each item of text in NVIVO is given a label (the title of the theme), and it is possible to run a query to draw out all text relating to that label. At first, there were a large number of labels that had been open coded (Burnard et al. 2008), but then it became possible to group them under 'umbrella' terms (Hycner 1985), resulting in an analytical framework. For example, the themes 'experiences of restraint' and 'alternatives to restraint' came under the umbrella term 'restraint' which in turn was grouped with 'seclusion' under the major theme 'coercive techniques' . Being able to read all of the information in a major theme was very useful, because this allowed an overview of what people were saying about the whole subject and the development of conceptual questions from these data.

Findings
The study produced many areas of analysis. For the purpose of this article, we focused on the main umbrella theme of 'restraint' . Subsequently, the following section discusses the emergent themes that were subsidiaries to this themegendered experiences of restraint, reasons for using restraint and alternatives to restraint.

Gendered experiences of restraint
Women service users on Unit C were described as using 'relational aggression' , which was considered to be more difficult to deal with than the type of aggression shown by male service users on the unit (see also Wilcox, Finlay, and Edmonds 2006;Williams, Scott, and Waterhouse 2001). This perceived difficulty legitimated the use of restraint with women, some of whom had been given a Borderline Personality Disorder diagnosis, a distinctly gendered label signifying extreme difficulties with relationships (Becker 1997). This sentiment was described by most of the staff who were interviewed; for example, John's comment: I think women are a bit more dramatic and they carry things on. Your male clients, they have an incident and it's finished, where the women could carry things on. (John, staff member) Despite the described low-level risk of physical aggression, restraint was being used regularly with the women who participated in this study. Not all women in the study had been physically restrained, and the ones who had not were very quick to point this out. Those who had not recently experienced restraint equally mentioned how long it had been since their last incident. All of the women commented that restraint did not help them to calm down: Louise's comments describe the humiliation involved in the embodied experience of restraint. All of the women mentioned that they did not like being restrained, particularly when other people were watching. The following quote demonstrates the gendered experience of restraint, and the feelings of powerlessness and indignity that can ensue: Kate's experience highlights the potential for abuse during and after incidents of restraint. Annie commented that restraint had also resulted in her body being exposed, and this made her feel that she was not safe on the unit: Most of the women who had been restrained reported that it felt painful (this was also found by Sequeira and Halstead 2001). Worryingly, it seems that the women were not given specific instructions about when restraint would be used. Helen, a woman detained in the unit, said: 'They don't tell you they're going to restrain you, you just know that if you do an act they're going to respond in that way. ' This statement implies that there is no warning prior to restraint, or guidance put in place that the women are made aware of.
There were some comments about the use of male staff members to carry out the restraint. Clearly, Annie felt violated by the use of men for restraining women, and she commented that incidents where restraint was used reduced when men were no longer involved: They've got it down to where I could only be restrained by females and that helped because I started getting restrained less more and when I did get restrained I wasn't fighting as much and I wasn't having injections as much, I wasn't going in seclusion as much. (Annie) These extracts illustrate the imbalance of power between the service users and staff coupled with the gendered concerns which particularly affect women. Furthermore, there is the potential for revictimisation of a sexual assault survivor, given the incidence of sexual violence in this population (McCarthy 1998). Staff member Marie described the use of male staff to restrain women: Marie: I think for a long time we relied on saying 'yeah, we need some men in' Researcher: To control?
Marie: Yeah, it feels very controlling, very authoritarian, 'yeah, we'll get the men in to sort you out' , it's ludicrous really.
Marie recognised the 'authoritarian' nature of this arrangement. The potential for retraumatisation when men are used in restraint is key: Ellie: I get worse when I'm lied on the floor because people with glasses -I mean men -I can't look at them Researcher: Okay Ellie: Because, I'll tell you. When I was in a care home in (place) a guy which was staff did something what weren't nice Three women service users mentioned that being restrained made them think about bad experiences from their pasts (see also Gallop et al. 1999;Fish and Culshaw 2005;Robins et al. 2005).
Staff member Steve acknowledged this imbalance of power and felt uncomfortable restraining women: It's never easy and as a man, sometimes having to restrain women is something that I wish I would never have to do. But I understand and accept that at times it is necessary, again either for the safety of the client or for other people really. Certainly I would always try and get female staff to restrain female clients first if that's a possibility. (Steve, staff member) It seems that Steve realised the potential for retraumatisation, yet felt that male staff restraining women was inevitable.

Reasons for using restraint
Despite the lack of guidance offered, women generally understood the reasons why restraint was used. For example: Researcher: Do you feel like you understood why they restrained you? However, as mentioned earlier by Helen, Annie pointed out that the occurrence of restraint was contingent on which staff were working at the time: Researcher: Are you saying that it's better if you know who's going to be restraining you?
Annie: No, you can never know who's going to restrain you because you don't really know if you're going to get restrained, because sometimes some people with restrain you and some people won't for certain things.
Researcher: So you don't even know why you get restrained.

Annie: Not always
Researcher: Do you get any warning or does anyone talk to you whilst -Annie: No they'll just do it until you calm down, sometimes they'll inject you sometimes they don't.
Annie's comments describe a worrying situation where one may not be aware of what circumstances will trigger a restraint, or even a forced sedative injection. Again, it is evidence that service users are lacking accessible information about restraint. Dawn, a staff member, acknowledged the physical discomfort of restraint from the point of view of staff, but like Steve described it as a necessary intervention: I think physical intervention is needed with our women and I don't think on our ward it's used too frequently. I think it's only used as a last resort and we will do generally anything to not restrain because nobody wants to be holding someone like that because it's the most uncomfortable thing in the world. (Dawn, staff member)

Alternatives to restraint
When asked how experiences of restraint could be improved, all of the women had some advice. Often women requested that staff talk to them and ask what is bothering them, possibly going with them to another room -for example: Brenda: Talking. Talking to us more … Talking to us more seeing why we're angry, letting us try to explain who we're angry with.
Researcher: Yeah, so you could get it out before?
Brenda: Yeah, let us try to explain.
If restraint has to be used, all of the women suggested that staff should be engaged with the person at all times: Researcher: Is it better if they're talking to you all the time?
Bonnie: Yes, not giving me no contact makes me worse. If they're holding me but not talking to me, I'm trying to talk to them and they don't talk to you back, that really does my head in and I go worse.
Talking with staff was important to all of the women, and some said that humour was a good way to calm them down. Humour is sometimes used as a way to de-escalate a situation before restraint is used; however, Ellie pointed out that staff laughing during restraint is hurtful: Ellie pointed out that staff offer her a PRN if she is unsettled, which is usually taken in pill form but in cases of extreme distress can be forcibly given as an injection as Annie described earlier. Women were requesting PRNs on the wards and being offered them, before the need for restraint. Some research studies argue that the use of PRN as a proactive method is much more preferable to women than more coercive methods (for example, Veltkamp et al. 2008). Ellie also said that she does not appreciate staff using humour in this situation. The use of humour for de-escalating aggression in intellectual disability services is discussed in the literature as effective on occasion (Duperouzel 2008), but it clearly is best used only by knowledgeable staff who have built up a relationship with the person, as Adele suggested here: It's about knowing your person again. There are some people for whom humour works and you can de-escalate a situation very quickly with humour. There are some people for humour is the big trigger and you wouldn't go there and you wouldn't make a joke of the situation, so it's about knowing the person that you're working with. In some instances you could remind them of positive things that are going to happen later on, and that's enough to make them stop and think about, 'Oh I don't really want to do this because I don't want to not be able to do that thing, ' and equally just reminding somebody of something may be the trigger to further escalation. So it really does come down to knowing the person to be perfectly honest. Which I think is so hard for people like bank staff 1 to know, and I'm sure that there are more incidents that happen with unfamiliar staff than with familiar staff. (Adele, staff member) Adele's comment makes an argument for the importance of a trusting relationship between staff and service users. The better this relationship is, the more the staff member knows how to work with the service user. Adele's comment demonstrates this when she points out that humour may curtail a situation for some people but not others. Her comments portray a certain amount of coercion, however, where the context and reasons for pending aggression are overlooked in favour of behavioural reinforcements for some service users. Evidently, incidents of aggression along with the corresponding use of coercive methods are reduced when staff/ service user relationships are at their best. Sometimes, ambivalence was shown towards good staff/service user relationships; staff showed concerns that women would engineer incidents of restraint so that staff engage with them, therefore sometimes staff were told not to interact with women during such situations -for example: Sometimes the guidelines are not to interact with them because they want that interaction and they want that touch. Some of them it's a sexual thing, and some want other women to be holding them all the time and talking to them. So then the guidelines say there's no interaction, make it quite boring for them, make it that it's not a good experience, in the fact that it's boring and it's not stimulating. Others that are very, very damaged, you would talk and just say things like you know, 'We're here, talk to us, you're safe' that kind of stuff. So it's all individual. (Karen, staff member) Karen making the point that women may be engineering aggression in order to gain some sort of engagement (see also Wilcox, Finlay, and Edmonds 2006) suggests that people are not receiving the responsiveness they need, or conversely that their anger is not considered legitimate and the context is denied. A particularly concerning matter was that there was no debrief after incidents of restraint (Wynn 2004). It could be argued that for the restrainer and the restrained to learn from the situation at hand, it would be beneficial for them to discuss it afterwards and tell each other how they felt, as Jackie suggested: Well it's surreal, I've always thought that if you were watching some of the things that happen on some of the flats here you would think 'what on earth is happening? How does that work that this restraint happens and then there isn't -not reconciliation -but there isn't a process that goes after? You don't want to reinforce something by having a physical intervention and then having a big confab or something, but some forum where you enable staff and clients to be able to talk with each other in an open and honest way, so that the staff aren't there as objects, or aren't seen as staff: 'They don't necessarily have feelings. They're there to do their job, they're there to sort this out and we are on the receiving end of this, But that makes us really angry, that we are. ' We need to acknowledge some of that stuff and share the responsibility because it's 'You made me feel like this, ' and, 'You did that' . So if you don't discuss it -to me communication's everything -it's an openness. (Jackie, staff member) Jackie's point about how life might look to an outsider indicates how conflict and resolution on the wards is very different from the outside world. Jackie was pointing to debriefing as a way of developing understanding between staff and service users, a way for everyone to understand how these situations make people feel. Yet her comment still reflects the dominant discourse of behavioural reinforcement. She went on to explore the whole concept of using coercive methods with the women: I'm really conflicted about it if I'm being honest, because I just think that, on the one hand I think these methods are probably necessary, on the other hand I just think once they happen it's like a line gets crossed and it confirms something to the individual woman client, that you wouldn't want to have them confirm to themselves. Like 'I'm so bad and out of control that I have to have this happen to me. ' (Jackie, staff member) Clearly, staff worry that these coercive methods are self-fulfilling due to their routine use, that the women will see these methods being used and accept that they are needed, and that this creates a downward spiral of events. Iona described her relationship with a service user in the past as particularly positive, and showed how this worked to eliminate incidents where restraint was needed: My attitude towards her was very upbeat and very OK. I was never negative towards her, if she couldn't do something I didn't say it in a negative way. And the relationship was very different between them two [other staff members] and her because they'd had all the aggression and I'd never had aggression [from her]. We'd have a laugh to be honest, we'd laugh and joke and I think there was an element of -what's the difference? (Iona, staff member) Iona's comment shows that aggression arises in a relational context, and importantly that a positive staff/client relationship can work to eliminate aggression and therefore the need for restraint.

Discussion
Coercive methods for behaviour management should be investigated using concepts of gender, power and control as well as disability (see also Griffith, Hutchinson, and Hastings 2013). Both groups of participants in this study pointed out the unpleasantness of restraint methods, and they emphasised the damage it could do to relationships. There were many opportunities being missed for discussion and resolution of conflict.
Gender was marked out in staff descriptions of women as exceptionally problematic people. This was expressed in judgements made by staff about aggression, in particular ideas about women's tendency towards relational aggression, and how much more difficult it is to deal with than the way male service users show aggression. The constitution of women as problematic on the unit set them apart from staff, 'othering' them, which in turn created conflict and expressions of distress. The intersection of disability and gender therefore created multiple situations involving devaluation and marginalisation (Aitken and Noble 2001).
Staff found expressions of distress extremely difficult and felt that they had no choice but to use coercive methods. This could be described as using physical security in place of relational security (Department of Health 2010), encouraging passivity and compliance as indicators or progression through services. Some literature suggests that there is a particular need for relational security for detained women Long, Fulton, and Hollin 2008). Relational security is defined by the Department of Health as 'the knowledge and understanding we have of a patient and of the environment; and the translation of that information into appropriate responses and care ' (2010, 5). Such a strategy relies on good relationships and is defined in terms of high levels of staffing, provision for staff and service users to spend time together -providing a balance between openness and intrusion -and high levels of trust (Parry-Crooke and Stafford 2009).
The importance of relationships is a significant thread running throughout the analysis. If staff know the women well, they are aware of their characteristics and how to be together in the safest possible way. This can work to reduce power imbalances when living in such close confinement (Knowles, Hearne, and Smith 2015). Supportive relationships were very important for women's progression in this research, and they also contributed to behavioural stability, the main quantifier of progression on the unit (Fish 2016;Travers and Reeves 2005). According to Herman's (1997) concept of recovery, the growth of good relationships between staff and service users can protect against the perceptions of repetition of past traumas as described so often in the literature about coercive methods (for example, Aitken and Noble 2001). Herman explains it as: '(A) shared understanding of the survivor's characteristic disturbances of relationship and the consequent risk of repeated victimization offers the best insurance against unwitting re-enactments of the original trauma in the therapeutic relationship ' (1997, 127).
It is accepted in policy that because women are a minority in locked wards, they are 'fitting in' to a structure which is more suited to men who have offended (for example, Corston 2007). It is further claimed that because of this, women are subjected to greater measures of security than they might need . Despite this recommendation, some researchers have demonstrated that women are detained within higher levels of physical security for less severe crimes than men, and certainly higher than is justified (Bland, Mezey, and Dolan 1999). It could be further argued that locked wards, like the rest of society, rely on a hegemonic (non-disabled) imaginary as the foundational principle of service delivery. An alternative, perspective would be designed through a deep and thorough engagement with people with intellectual disabilities, and would enshrine the principles of maximum respect, effective communication centred on the provision of accessible information, and the need for disability accommodations throughout every practice and process.
All of the participants in this study recognised that the nature of women's experiences of trauma, emotional expression, relationships and restraint were profoundly different from men's. These gendered experiences are often under-appreciated in the literature about coercive methods. We argue that the use of restraint should be reframed, taking into account gender as well as disability dynamic. One outcome of such reframing could be to stimulate a broader social policy discussion on the ways in which gender-informed research can challenge existing practices and suggest more gender-aware responses. Because disability and gender co-exist in intimately connected ways, such discussions should draw on a feminist disability studies perspective where this intersection is thoroughly emphasised (Carlson 2001). The harms experienced by all people with intellectual disabilities deserve greater attention, but given that these women have experiences of emotional and physical abuse which are related to their gender, there is a urgent need to explore these disparities in the light of gender-informed well as disability studies research.
Further, the experiences of adverse effects related to restraint -and indeed their prevention -should be explored through a gender-sensitive analysis of the alternatives to restraint. In this study, women suggested ways of working together which would reduce the need for restraint, such as allowing them to articulate anger and the reasons for it, and strengthening of the therapeutic relationshipboth of these rely on changing the power dynamic. A disability studies perspective would also demonstrate the ways in which the experiences of people with learning disabilities are relatively discounted in planning social services, treatments, institutions and social policy.
The unique approach adopted in this study, engaging with both feminist and disability studies approaches, opens the door to exploring new responses in policy and practice. The collective wisdom of people with intellectual disabilities (like that of other disabled people) could be a real source of collective strength and empowerment. Advancing changes in power relationships would involve challenging established power relationships in treatment, and challenging attitudes towards women with intellectual disabilities in particular. So much of the harm to the women's self-image, relationships, life opportunities and self-determination has been fostered in environments that involve experiences of sexism and disablism. These broader dynamics of social exclusion and disabling barriers also need to be challenged. Restraint is often the last-ditch temporary solution to a crisis; the long-term goal must be to work inside and outside institutions to prevent so many people with intellectual disabilities being put into such precarious positions that they fall into crisis.
To create outcomes which have a positive transformative potential for women with intellectual disabilities, we need to recognise the need for advocacy, support for appropriate alternatives to restraints, and to fund situations where women with intellectual disabilities who have experienced these institutional practices can mentor their peers. This would be only one (albeit profound) change in the ways in which 'challenging behaviour' is conceived -as a catalyst for reflectionbut it may be one sign that policies and practices in this area are heading in the right direction. Note 1. Bank staff are provided by an agency and are used on an ad-hoc basis to cover periods of short-staffing. They may not have any prior experience of working with the service users.