Barnahus for Adults? Reinterpreting the Barnahus Model to Accommodate Adult Victims of Domestic Violence

The key purpose of this chapter is to identify challenges in terms of reinterpreting the Barnahus model to accommodate adult victims of domestic violence. The analysis is based on data from the first part of an ongoing process evaluation of the November project in Oslo, Norway, which is set in a police station and aspires to be a Barnahus for adults. Drawing on both the Norwegian Barnahus model and a Swedish multi-agency model for adult and child victims of domestic violence, this project constitutes a complex translation process. As two related but disparate source models are brought together and translated across both national borders and age divides, there should be an acute awareness of the particular recipient context, especially in terms of existing collaborative structures and unmet needs. The analysis is relevant for countries that are considering the implementation of the Nordic Barnahus model because it highlights the importance of a detailed analysis of the local institutional landscape.

this Norwegian project has developed during the early phase of its implementation. What kind of collaborative model does November set out to be, and how are these ideas negotiated and brought into practice? The key purpose of this chapter is to identify challenges in terms of reinterpreting the Barnahus model to accommodate adult victims of domestic violence in the current landscape of policies and measures against domestic violence in Norway. This analysis of Project November may highlight important issues to consider for the implementation of other types of collaborative models, such as the Barnahus model. The discussion will be linked to theoretical concepts on multi-agency and multi-professional collaboration (Atkinson et al. 2007).
The chapter is based on results from the first part of an ongoing process evaluation of the November project. 1 Empirically, we draw on different types of data: Written material that includes initial project plans describing the model and process documents from the implementation phase and qualitative interviews with 21 key stakeholders and operative personnel within November. The interview material is central to our analysis but as anonymisation is complicated, we will refrain from direct quotations.

Conceptual Framework
In both Norway and Sweden, a key motivation for the original Barnahus models as well as the subsequent models for adults, November and Karin, was to facilitate coordination between the different agencies and services involved in police-reported incidents of violence and sexual abuse. The concept of inter-or multi-agency collaboration is often used to explain such models (for instance, Johansson 2011Johansson , 2012; however, the meaning of multi-agency collaboration is rarely spelled out. Several scholars point to the confusing terminology used to describe collaborative arrangements and multi-agency activity, which makes classification and comparison difficult (Atkinson et al. 2007;Percy-Smith 2006;Blacklock and Phillips 2015). Different researchers use words such as partnership, joined-up, coordinated, integrated, co-located, and more for such arrangements. We also note that the concepts linked to multi-agency collaboration are all positively connoted. Multi-agency work can, of course, result in professional competition and conflict, and the analysis also needs to be sensitive to this (see Johansson,Chap. 16 in this book). Atkinson et al.'s (2007) differentiation between multi-agency and multi-professional working is fruitful for our purpose. According to their definitions, multi-agency or cross-agency working represents situations where professionals from more than one agency work together, and service is provided by agencies acting in concert and drawing on pooled resources or pooled budgets. Multi-professional or multidisciplinary working denotes situations where staff of different professions, background and training work together within the same agency. Our question then relates to how the Barnahus model, Project Karin and Project November may be classified according to these concepts.
Given that Project November is inspired by two other collaborative models, Barnahus and Project Karin, the implementation of November can be seen as an act of translation (cf. Johansson 2012). It is an example of how organisational ideas circulate or travel between actors and places. Czarniawska and colleagues designated the study of such processes as the "sociology of translation" (in Røvik 2016, 291). A key understanding from this perspective is that ideas are transformed while being transferred from one organisational context to another. From this perspective, Barnahus and Karin represent "source" models, meaning organisations that are "performing a practice and possessing knowledge about it that someone attempts to transfer to another organisation (recipient unit)" (Røvik 2016, 290). Project November represents the recipient unit in this framework. Røvik points to the need for process data to capture what happens to ideas after their formal adoption, suggesting that ideas are shaped and reshaped as part of implementation processes. Local translations often lead to the emergence of new versions, and significant variation in structures, routines and practices (Røvik 2016, 291). In terms of Project November, such variation could be expected given that the project was designated from the beginning as a pilot version of two different, albeit linked, institutional models for inter-agency working.
The analysis in this chapter will focus on two interrelated processes of reinterpretation: the translation of two related, but in some respects different, source models for multi-agency collaboration, and the translation of such models across different national contexts, as well as different target groups. First, however, the next section briefly presents the two source models, Barnahus and the Karin, followed by an outline of the Norwegian context.

Background The Source Models: Barnahus and Project Karin
As noted, Project November originates in the current national action plan against domestic violence (Ministry of Justice and Public Security 2013). In both the plan as well as in the main project documents, the project is described as a "Barnahus for adults" on the one hand and as a Norwegian pilot version of the Karin model on the other.
As explained in more depth in other chapters of this book, the remit of a Norwegian Barnahus is to facilitate and coordinate the process of forensic documentation in cases where there is a suspicion that a child has witnessed or been the victim of violence or sexual abuse. A core idea is to bring the relevant professionals together under one roof and in a child-friendly setting. Other core activities include the child investigative interview performed by specially trained police personnel and medical examinations by specialist doctors and nurses from cooperating paediatric departments. Barnahus staff coordinate the process and provide follow-up/treatment on a short-term basis, and work with child welfare services to ensure seamless referral to local agencies. The Norwegian Barnahus model includes a permanent staff of psychologists and social workers, and police investigators, medical staff and representatives of child welfare services and other agencies are present on a case-by-case basis. This is in contrast to the Swedish Barnahus model, where child welfare services are part of the Barnahus organisation (cf. Bakketeig,Chap. 13). Through the use of link workers, it is nevertheless a coordinating facility. The Swedish Barnahus thus qualifies as a multi-agency model in Atkinson et al.'s (2007) understanding of this as more than one agency working together with pooled budgets. The Norwegian Barnahus, on the other hand, cannot be considered a fully developed multi-agency model in this sense. Multi-professional working seems to be a more applicable categorisation. Still, the model involves some sense of bringing together agencies through the provision of a common physical space and collaborative routines. As pointed out in the introductory chapter and by Johansson in Chap. 12, the Norwegian Barnahus organisation is perhaps most accurately described as a multi-professional competence centre. Institutionally and in terms of budgets, Norwegian Barnahus belong to the relevant police districts and are most often located in separate buildings, although a few are part of a police station.
Project Karin in Malmö Sweden on the other hand has a broader remit than the Barnahus, in terms of both target groups and organisational structure. It is a police station that specialises in helping women, men and children who are the victims of violence or sexual abuse by a closely related person. The project also supports children who have witnessed violence, and it can assist people who have been reported to the police (suspected perpetrators). Notably, in terms of being a collaborative model, Karin is the result of a joining of forces by two specialist units working with domestic violence in the municipality of Malmö: one in social services and one in the police force. As social workers and criminal investigators respectively, they have retained their affiliation with their mother institutions, but at Karin they collaborate closely and liaise with other organisations and public authorities. Shelter services for adults and a children's service similar to Barnahus are also organised in conjunction with Project Karin (BRÅ 2013). The physical environment is designed to create an atmosphere of security and comfort, much like the Barnahus (see Stefansen,Chap. 2). In other words, Karin is a more comprehensive, and in some respects more self-sufficient, model than the Barnahus in terms of target groups, facilities and services offered in one location. As personnel are employed in two different agencies (social services and police), it qualifies as a multi-agency model in Atkinson et al.'s (2007) terminology.

The Norwegian Context: Public Policies and Services in the Field of Domestic Violence
Following years of awareness raising and capacity building, the current government action plan (2014)(2015)(2016)(2017) confirms that domestic violence has become a mainstream policy issue and that service provisions are in place. According to the plan, the main challenge is now to improve coordination between services and agencies, creating a more coherent and efficient chain of services. Testing a multi-agency model inspired both by Project Karin in Sweden and by Norwegian Barnahus is among the measures introduced in the plan.
Other previous and recent reforms include a system of special domestic violence coordinators in the police districts (from 2002), the introduction of Barnahus (from 2007), municipal action plans and a law on shelters for victims of domestic violence (implemented in 2010, see Bakketeig et al. 2014). The shelter law is particularly relevant to our analysis of Project November as a translation of collaborative models in a new context, as from 2010 it is a statutory municipal duty in Norway to provide shelter services to victims of domestic violence, irrespective of gender. This special law goes beyond shelter services in the traditional meaning of a safe house for victims in need of immediate protection and peer support, however (Stefansen 2006). In fact, the law sets out a comprehensive range of duties for local authorities, including the following services, all free of charge: a 24-h telephone helpline, shelter or equivalent safe, temporary accommodation, a day service including support and counselling, and assistance during the re-establishment phase. The latter includes guidance in establishing contact with social services and other relevant agencies. In providing such services, the law obliges municipal authorities to pay special attention to children's needs and, most notably in the context of this chapter, to coordinate different services. This coordination duty is closely connected to municipal action plans that are not mandatory, but strongly encouraged by central authorities and supported by guidelines. In sum, these policies mean that the role of Norwegian shelters is considerably different from shelters in Sweden and other European countries, where they more often belong to the third sector. In a Norwegian context, domestic violence shelters have become part of the public sector on a par with other contracted services. In fact, the majority of shelters are today run by municipalities (Bakketeig et al. 2014).
Major achievements notwithstanding, there is considerable diversity among public agencies regarding domestic violence competence and priorities. While the police force has made considerable progress, the health and social services are often claimed to be less informed about and sensitive to issues of domestic violence (Grøvdal et al. 2014). Shelters vary, however, in their evaluation of collaboration with social services. Collaboration is rated as good in some municipalities and as difficult in others (Bakketeig et al. 2014). All in all Project November was introduced into a complex institutional landscape of already existing service provisions for adult victims of domestic violence. A key question we address in this chapter involves the extent to which this is recognised in the translation and implementation process of the model.

Analysis Project November
Initiated as a measure in the national action plan against domestic violence, the overall aim of Project November is to provide better, more holistic and coordinated services to adult female and male victims of domestic violence. From the ministry's point of view, the idea was to bring agencies together "under one roof ", allowing victims to access services through "one door" (see Chap. 1), or as in the wording of the plan, in the same locality: Establishing "Karin" (Barnahus for adults) as a pilot project. A collaborative project equivalent to Karin in Malmø, where the police and other agencies offer assistance to victims of violence in the same locality, will be tried out. (The Ministry of Justice and Public Security, 2013, p. 28) In terms of aims, there are clear resemblances to the Barnahus model and the discussions prior to implementation. A multi-agency project group worked together to design the model in the preliminary phase, but the project was placed firmly within the police sector with only limited participation from the municipality in the steering group.
Project November is set in Stovner police station, which houses the centralised family domestic violence unit of the Oslo police district. It is located in a designated part of the police station building that has been designed specifically for this purpose, including a room dedicated to investigative interviews with audiovisual equipment and with an adjacent co-hearing room. Aesthetically, the furniture and interior decoration have much in common with the Oslo Barnahus (see Chap. 2) and Karin, described above, and it is very different from the other parts of the police station.
So far, the November staff consists of seven people who have been recruited specifically for the project. All are employed in the police district. The staff is divided into two teams: a police team with two police specialists in domestic violence risk analysis and a psycho-social team with two social workers (one clinical specialist and one with experience in the social services) and a psychologist. It should be noted that the Oslo police distinguish between risk analysis on the one hand and criminal investigation on the other, in order to cultivate risk management as separate expertise. In contrast to Karin then, police investigators are not included as such, but the November staff cooperate with investigators and prosecutors from the regular police station. The project is headed by a police leader, assisted by the leader of the psychosocial team.
The psycho-social staff partly assist the police and partly work on their own initiative. In particular, the psychologists, but also social workers, assist the police in risk assessments of perpetrators. In other words, one added value of November is directly related to improving police work. The psycho-social support that was previously offered to some extent by the police is now taken care of by specially trained social workers. In this sense, Project November relieves the police of their workload and thus contributes to freeing more time for investigative and preventive work.
In addition to assisting the police team in risk assessments, the psycho-social team engages in a range of victim support and counselling, including outreach work. They provide psycho-social support, stabilise victims in acute situations, inform them about the police process, give advice on how to approach other agencies, assist and follow them to the social services and housing offices, and in short do advocacy work in line with common definitions. Hoyle and Palmer (2014, 14), for instance, offer this definition of advocates: "Advocates 'direct, guide, and support battered women while confronting and challenging obstacles to their safety (Shepard 1999, 115)'. They counsel, provide access to resources, represent victims in other institutional settings, such as the court, but more importantly they help women to better understand their options and how to make choices." In this sense, the advocacy work of the November staff resembles the advocacy work of the staff at domestic violence shelters (see Bakketeig et al. 2014). A central concern for the November staff is to motivate victims to file a police report and to contribute to the investigation and punitive process. Victims who are not yet part of a criminal investigation can obtain information and motivational support.
If cases involve children, either as victims of, or witnesses to, violence, they are referred to the Oslo Barnahus according to ordinary procedures. Perpetrators, on the other hand, are part of Project November's target group, in contrast to Norwegian Barnahus but similarly to Project Karin. The tasks relating to perpetrators include motivating them to seek treatment, typically in terms of anger management and treatment for alcohol and substance abuse. The psychologist and the clinical social worker also offer short-term treatment sessions, and the psychologist refers victims to the psychiatric services and perpetrators to specialist therapy for perpetrators of domestic violence. 2 In sum, there is no doubt that November represents a more extensive support service for victims and perpetrators, than the ordinary police station, and in particular in terms of social work. One example described in an interview at November is that of a woman whose former partner had been violent and had also taken out loans in her name. A police report was filed, and he had violated his restraining orders. The social worker at November spent a considerable amount of time asking different creditors to freeze these loans pending the police investigation. There was also a general consensus among the staff that social services were particularly important but also reluctant to take on domestic violence as part of their expertise and remit, much in line with Grøvdal et al.'s (2014) findings.
Victims of domestic violence are also offered more qualified psychosocial support in the context of a police investigation and, to some extent, beyond. Our interviewees from the psycho-social team stressed the benefits of having more time at their disposal. They could offer repeat contact at the same place, instead of having to refer victims to other agencies and risk them dropping out on the way.
Overall, the November staff has a wider remit than the ordinary domestic violence police in that they are less restricted by the priority of severe and police-reported cases. The staff explained that cases which would otherwise be thought to fall outside the police priorities were "picked up" by them, including what are typically termed "difficult" or "lost" cases, such as those involving episodic and bilateral violence patterns in conjunction with alcohol abuse, for example.
In their typology of different collaborative models, Atkinson et al. (2007) differentiate between three dimensions of multi-agency activity: organisation, joint investment and integration. In terms of organisation, Project November is funded by one agency, the Ministry of Justice and Public Security, through the Police Directorate, and it is owned and managed by the same source. In this sense, the project is thus not multiagency. In terms of joint investment, there have so far been limited efforts to create links with and to commit agencies outside the police. Integration is low in relation to other agencies but high within the project itself. The multi-professional staff work closely together on a day-today basis involving a mutual exchange of knowledge and skills. There is a flexible division of tasks, such as when it comes to the role as the main contact for a victim or perpetrator, and professional roles may overlap to a certain extent.
In short, the ambition of offering victims one door into the coordinated services of multiple agencies that are co-located under one roof has not (yet) materialised. November is not a multi-agency model in the sense of bringing agencies together and pooling resources. In Atkinson et al.'s (2007) terms, it is more a case of multi-professional or multidisciplinary collaboration. At this stage of the implementation, there is thus a question of whether and how to proceed in terms of multiagency integration. It seems that Project November will have to choose between moving towards a more truly integrated inter-agency model and settling with being a multi-professional support centre for adults involved in police cases about domestic violence, be they criminal cases or risk management cases.
In Karin, by contrast, social workers are employed in the municipal social service. Moreover, Karin offers a more extensive service in that they are co-located with a shelter and a version of Barnahus. As an inter-agency model, November thus seems to be more similar to the Norwegian Barnahus than to the Karin Project in Malmö.

Key Challenges for a "Barnahus for Adults"
So far then, we have established that Project November is a multi-professional support centre for adult victims and perpetrators of domestic violence in a police context. This model is still a preliminary outcome of a translation process in the making; however, the first phase of implementation has brought some key issues of general interest for the translation of practices that are complex, partly tacit and embedded in local structures. In the following discussion, we will focus on two sets of interrelated challenges.

Translating Across National Borders: Issues of Particular Contexts
Scholarship on multi-agency working stresses the need to be clear about roles and to conduct a needs analysis (Atkinson et al. 2007, 44-45). Analyses of needs should be context specific and may include questions such as: What is the particular need or gap to be filled? Who are the potential collaborating partners and are there any competitors or overlaps?
In the case of Karin, that particular model grew out of the process of developing a municipal action plan against domestic violence (BRÅ 2013). The Malmö municipality already had a specialised domestic violence police station on the one hand and a special unit for domestic violence within the social services on the other. Establishing Karin involved bringing these two specialist parts of two agencies together to create a multiagency facility. In Norway, the Barnahus model was introduced by national authorities, but after much lobbying and pressure from NGOs, politicians and professionals Stefansen et al. 2012;Chap. 16 in this book). In the policy field of violence and abuse against children, there seems to have been a consensus that a coordinating mechanism was lacking. Barnahus thus filled a void, in terms of both providing enhanced and specialised police procedures and functioning as a multiagency coordinating mechanism for the follow-up phase after the police interview.
The process in November has been more top-down and may prove to become more contentious. As already noted, the initiative came from the Ministry of Justice and Public Security, and the Oslo municipality was not involved until later and so far only to a limited extent. Norwegian municipalities are strongly encouraged to develop action plans against domestic violence as a key coordinating tool. In Oslo's case, there is both an overall plan for the city in general and local plans for all fifteen boroughs, reflecting the considerable degree of decentralised self-government in the Norwegian municipal system. The overall municipal action plan mentions November only briefly, under the heading "Cooperation with the police" (Oslo Municipality 2014, 8), and notes a health agency as the main coordinating agency for adult domestic violence cases, the Emergency Social and Medical Team, which is part of the Oslo municipal emergency ward (Sosial og ambulant vakttjeneste, Oslo legevakt) (Ibid., page 9). At the central municipal level in Oslo, there is, however, another specialist agency, the Oslo domestic violence shelter. Both the shelter and the Emergency Social and Medical team are 24/7 services and both are, to some extent, given coordinating roles in relation to adult victims of domestic violence. In other words, Project November was introduced into a local context which already had certain structures for coordination and where there seemed to be some degree of tension built into these structures. To illustrate the significance of such pre-existing structural conditions, we will use the shelter as an example.
The Oslo domestic violence shelter was established in 1978 as a refuge run by volunteers and based on the idea of peer support, but later developed into a professional support centre for domestic violence victims (Bakketeig et al. 2014;Jonassen and Stefansen 2003;Laugerud 2014;Stefansen 2006). Today, it consists of a residential section for women and an outpatient section providing counselling, advocacy and courses as well as a 24/7 telephone line. There are also three special accommodation units: one for young women exposed to honour-based violence, one for female victims of trafficking and a separate unit for male victims (Bakketeig et al. 2014). The shelter is in charge of their own intake procedure and has contracted working agreements with the municipal social services in all boroughs. According to these agreements, each social service office assigns a special liaison person to cooperate with the shelter. This coordinating arrangement follows from the shelter's role as Oslo's statutory municipal shelter provision, based on a 10-year contract in keeping with the shelter law noted above.
This role as a comprehensive public service provider is still relatively new, however, and in the shelter's own opinion, they have not yet been fully acknowledged by other municipal agencies. There is, for instance, limited reference to the shelter in the municipal action plan. In fact, there seems to be a certain tension between the role assigned to Oslo municipal emergency ward and the statutory role of the shelter. Moreover, there may seem to be some lingering prejudice associated with the shelter's NGO and self-help past, as well as a certain understanding that some women do not want to use the shelter.
Overall, these particular circumstances mean that the establishing of November coincides with a somewhat fluid situation when it comes to the roles of already existing specialist agencies in Oslo and to the question of coordination. Such structural circumstances could imply that existing agencies may see November as a competitor and not (only) as a partner. When it comes to the more generalist agencies, such as the social services, the structure of relevant municipal and state partners for November is a conglomerate of centralised, semi-centralised and decentralised agencies. Within November, there is an ambition to establish cooperation agreements with the social services and housing that belong to the borough-based services. As noted, social services are considered especially important, but difficult to include into the chain of support for domestic violence victims. Compared to Barnahus in Norway, where the child welfare service is the equivalent agency, the duties of social services in connection with (adult) domestic violence seem to be less developed and acknowledged by the agency itself.
In Karin, the social services are an integrated part of the collaborative model. Social workers were drawn from, and have retained their employment in, social services. It is notable, however, that they do not have the authority to make decisions about social benefits or any other economic support. To claim such support, Karin's clients still have to approach their local social service office, albeit with the help from Karin's social workers. To this extent, November and Karin might not be so different after all, although November's social workers are part of the police, while Karin's liaise with colleagues. Nevertheless, the evaluators of the Karin model point to this particular element as a major limitation of such police-based models centred on the judicial process, which is only one part of the victim's process towards a life free of violence. They stress the need to develop additional, well-functioning systems that take over when the judicial process is finished. This includes, for instance, housing, economy, and the social area. The social services need to cultivate the kind of skills and organisation that, in addition to initial support, can provide help based on a decision to offer public assistance; otherwise, there is a risk that the efforts of Karin-and other similar organisations-will be a temporary relief but nothing more (BRÅ 2013, 31).
A central challenge, therefore, in the next stage of implementation, is whether November can develop new and innovative working relationships with the social services. What should the aim and division of work for such cooperation be, in view of the perceived reluctant attitude of the social services in cases of violence? Will the social services welcome cooperation as a way of being relieved of "difficult" cases, or will more cooperation result in more competent and adequate social services for this group?

Translating Across Age: Issues of Autonomy
The demarcation of both Norwegian Barnahus and Project Karin is in police-reported cases, while November has taken a more open approach and includes cases that are not yet reported and perhaps will not be. The issue of police reporting is not just a matter of priority, however. It evokes a more general question of victim autonomy that is central to the difference between a multi-agency model for children compared to one for adults. To put it very briefly, as far as children are concerned, the suspicion of violence or sexual abuse would always result in some kind of police involvement. In the case of domestic violence against adults, the question of police involvement is more open and contested.
Arguments for collaborative arrangements in tackling domestic violence against adults often refer to the challenge of breaking the cycle of abuse. It is argued that prosecution needs to be supplemented by other types of support in order to create sustainable change in the lives of victims and perpetrators lives. Adding civil to criminal remedies is both an aim in itself and a tool to increase conviction rates. Some see psychosocial support primarily as a tool to improve and enhance the punitive process, by motivating victims to cooperate with the police, giving evidence and remaining "cooperative" witnesses. Others argue that psychosocial support is needed to improve victim autonomy and choice about how they want to live their lives. The latter set of arguments is often framed in terms of victim empowerment (Hoyle and Palmer 2014). In discussing what an empowerment approach means in a domestic violence context, Hoyle and Palmer define it as helping clients to understand the choices that are available to them and to make informed choices, as well as supporting them in the choices they make. According to Hoyle and Palmer, empowerment thinking is still relatively rare within the criminal justice system. They claim that focus is more on "providing services in the interest of the victim", than engaging with victims about what they see as the best solution to their problems (Hoyle and Palmer 2014, 3).
Such issues of autonomy and choice were also raised in the interviews with November staff. One of the informants pointed to differences between the police and the psycho-social professions when it comes to approaching so-called reluctant victims. She explained that the police tend to be action oriented and impatient, sometimes to the extent that they want to make choices on behalf of the victim. In contrast, her psycho-social training inclined her to give the victim more time. She said that she encouraged her police colleagues to think of tackling witness reluctance as planting a seed. By working with the person over time, their wish to participate in the investigation and to leave the abuser will often become stronger, she claimed.
Reinterpreting Barnahus to accommodate adult victims should consider how victim autonomy may represent challenges that are different from those at Barnahus. As noted, suspicions of violence against children will involve a more clear-cut procedure of criminal investigation and police involvement. Of course there are also empowerment and autonomy issues when children are victims; however, agencies have a more unambiguous duty to intervene. The agencies are also not the same for child cases and adult cases, and their remits reflect the different status of children and adults as victims. For instance, child welfare services are not conditional in the same way as the economic support and housing benefits issued by the social services for adults. Such provisions are crucial elements when adult victims consider their options in relation to the perpetrator, to stay or to leave. Social workers and the psychologist may motivate and work with the victim to overcome psycho-social barriers to reporting the violence but this choice is also contingent on more concrete alternatives.
The view of autonomy could also imply that, in contrast to child victims, adult victims may want to choose between agencies, depending on their specific situation. Typically, some women prefer to seek refuge or advice at a shelter but refrain from involving the police. As the question of punitive intervention in domestic violence cases is still an ambivalent issue, it may thus be argued that issues of autonomy and choice favour a range of services and agencies that do not necessarily merge into one coordinated whole.

Discussion
It seems that November was based on a rather generalised notion of the problem of fragmentation and a correspondingly generalised idea of what was needed as a solution. The Ministry of Justice and Public Security decided that the project be housed in a police station in Oslo, with very little analysis of needs or the consequences in the particular local context. In other words, it seems like a rather top-down initiative that was nevertheless given an enthusiastic welcome by police leadership, both at district and at police station level. The involvement of agencies outside the police, including the central municipal authorities and boroughs, as such, has been limited, however.
In taking the pilot project forward, it seems that a project like November has a choice between two models: the present multidisciplinary police station or a more truly multi-agency organisation. In the case of the latter, there seem to be two ways of achieving such collaborative arrangements. The most ambitious is a co-located facility where specially dedicated personnel from several agencies work together in the same physical space, offering victims "one door" in literal terms. This would take the model in the direction of Karin but ideally add decisionmaking authority for social benefit and housing issues. A less ambitious solution would be that November personnel cooperate with specially assigned personnel in other agencies who mainly remain physically in their localities but work closely with November on a case-by-case basis, as in the Barnahus model.
It appears that the more ambitious the multi-agency intentions, the stronger the need for an assessment of the particular local context in terms of existing collaborative structures and unmet needs. In the case of November, on the one hand there probably is a real void when it comes to coordination mechanisms and structures, in particular for bringing the social services on board. On the other hand, the void is not clearly defined and there are already other specialist agencies geared towards trying to fill it. Accordingly, if November opts for a fully fledged interagency model, there is a risk of duplication and competition.
Moving towards a multi-agency model would also raise the question of location. Why a police station? When the Norwegian Barnahus was established, as part of the police districts, a point was made about locating it separately from ordinary police work and at a central, convenient place for users (cf. Chap. 2). Barnahus is already an established competent organisation specialising in domestic violence cases and multiagency work. One could envisage the existing Barnahus being extended and adapted to include adult victims, although that would entail a much broader regional scope than planned for November. On the other hand, as noted there are already considerable elements of coordination within the mandates of both the Oslo domestic violence shelter and the Oslo municipal emergency ward.
Our analysis of this reinterpretation process shows how the outcome is determined by a translation in two dimensions. The idea of a Barnahus for adults draws on two quite divergent inspirational sources, consisting of a range of elements that are included or excluded, reproduced or modified as they are adapted in the recipient unit. Some combinations of elements may seem more feasible or difficult than others. For instance, issues of autonomy and choice pose different challenges in a model targeted at adults, compared to one developed for child victims, and vice versa. On a more general note, it seems necessary to specify the type of collaborative model that is envisaged, in terms of the scope and level of collaboration. For instance, there is a choice between a multidisciplinary model within one agency or sector and a more comprehensive inter-agency model spanning multiple sectors. In particular, the latter will require extensive multi-sector support from the top and concrete adaptation on the ground. This actualises the second dimension, the varied contexts that surround and impact the source and recipient models. There should be an acute awareness towards the particular recipient context, especially in terms of existing collaborative structures and unmet needs; otherwise, introducing a new multi-agency framework without analysing the local setting may end up as duplicating and competing effort.

Notes
1. The process evaluation is conducted within the framework of the Domestic Violence Research Programme (2014-2019) at NOVA. The programme is financed by the Norwegian Ministry of Justice and Public Security. 2. Oslo has a specialist treatment centre for perpetrators of interpersonal violence, Alternative to Violence.