Introduction

The co-occurrence of domestic or family violence (DFV) and child protection concerns has long been established in the literature (Hartley 2002). Exposure to DFV has deleterious effects on children’s health, development and mental wellbeing, with potential for long-term behavioural impacts into adulthood (Holt et al. 2008; Horton et al. 2014). Children are not only present as affected-observers; DFV can escalate to direct involvement in the violence, which carries with it the risk of child homicide (Reif and Jaffe 2019). The risk of serious injury or death makes managing high-risk cases of DFV a key focus for practitioners working with affected families (Juodis et al. 2014). Despite the gravity for women and children, there is a paucity of research on safely and effectively working with families with co-occurring child protection and DFV concerns (Humphreys and Campo 2017), with less attention on managing high-risk cases. Such research can benefit practitioners in child protection services, in particular, but also those working in any of the involved stakeholder services collaborating in the risk management of families. This article examines workers’ reflections and observations of practice in high-risk cases at the intersection of DFV and child protection in Queensland, Australia. A total of 15 participants, all of whom were senior practitioners, team leaders or managers from statutory child protection, non-statutory family services, DFV services, police and probation and parole participated in the research. The following section of this paper reviews the literature on interventions in DFV where there are child protection concerns, including multi-agency responses, which provides a detailed background to the study.

Multi-Agency Responses to Risk

The multi-agency approach, which has its origins in the Duluth model’s response to DFV, is increasingly common practice in child protection cases, as the needs of all family members cannot be met by any one service (Humphreys and Campo 2017; Stanley and Humphreys 2014). Cases referred for multi-agency or integrated responses are generally those assessed as high risk (O'Leary, Young, Wilde & Tsantefski 2018; McLaughlin et al. 2018). High-risk cases can make collaboration amongst agencies more challenging in relation to both victim/survivor and worker safety (Healey et al. 2018). A review of recommendations produced by domestic violence fatality reviews found that lack of information sharing or referrals beyond agencies, and poor service co-ordination were not only challenges agencies working with survivors to prevent homicide faced, they contributed to the risk of homicide (Reif and Jaffe 2019).

To manage these cases, service providers need to identify which perpetrators display psychopathic traits and assess the potential for lethality (Juodis et al. 2014). However, there are differences in the ways in which workers from disparate agencies, such as domestic violence, justice, and child protection services assess and manage risk, hold perpetrators of violence to account, and work to keep women and children safe, adding further complexity to collaboration (Healey et al. 2018; Stanley et al. 2011; Wilson et al. 2015). Multi-agency risk assessment conferences, such as the MARAC in the U.K., have been found to reduce recidivism among high-risk cases (Robinson and Tregidga 2007). Specific DFV homicide risk-assessment tools, such as the lethality screen, have also been developed for collaborative practice and are routinely used to enhance multi-agency responses in the UK (Graham et al. 2019), Australia (Lauria, McEwan, Luebbers, Simmons & Ogloff 2017) and Canada (Olver and Jung 2017), among other jurisdictions. However, Graham et al.’s (2019) systematic review of risk assessment tools developed to assess risk in relation to DFV, and DFV homicide, established that more attention is needed to determine the feasibility of these tools in practice settings. For example, Australian researchers found that some commonly used tools, such as the Spousal Assault Risk Assessment (SARA), the Ontario Domestic Assault Risk Assessment (ODARA) and the Family Violence Investigative Report (FVIR) are most predictive of risk when used in tandem (Olver and Jung 2017).

Partnering with Victim-Survivors in Domestic Violence Cases

Building an alliance with, or ‘partnering’, with adult victim/survivors of domestic violence has been promoted as a means of improving service engagement and increasing the safety of adult and child survivors (Mandel 2014). Working closely with victim/survivors provides vital information on perpetrators’ patterns of behaviour and victim/survivors’ own view of the level of risk, thereby enhancing risk assessment and management (Mandel 2014; Stanley and Humphreys 2014). In working with women in high risk situations, practitioners need to assess the life-generated risks which may keep them in their circumstances: financial constraints; housing; stability of children’s schooling; and the ability to meet basic needs can all be deterrents to leaving (Humphreys and Campo 2017). Children exposed to DFV also want their experiences acknowledged and included by service providers (Holt et al. 2008). Scott et al. (2020) note that inclusion of children’s experiences by frontline justice personnel improves information gathering, especially in relation to coercive control, but that justice services have limited interactions with children, positioning them as observers rather than victim/survivors of DFV. Further, the focus on children across multi-agency responses can enhance monitoring of high-risk fathers, especially post-separation and in custody disputes (Scott et al. 2020).

Safety Planning in High-Risk Domestic Violence Cases

Safety planning with victims/survivors is a key strategy in managing risk; however, its effectiveness relies on the availability of community and personal resources, efficacious collaboration of services, including non-specialist services, and the overcoming of potential barriers in implementing the plan (Campbell 2004; Logan and Walker 2018; Murray et al. 2015). Limitations of safety planning include: inconsistency across services; not addressing emotional safety or any co-occurring mental health or substance use issues; neglecting the long-term needs of victims/survivors; and failing to hold perpetrators to account for their behaviour (Murray et al. 2015). Safety planning is, nevertheless, an effective means of partnering with victims/survivors that helps to avoid interagency collaboration that privileges the professional view over victims/survivors’ own perspectives (Jenney et al. 2014; Stanley and Humphreys 2014).

Engaging Fathers Who Use Violence

Along with the involvement of an array of stakeholders, contemporary approaches in child protection call for a whole of family response, with a shift away from working solely with women and children to include working with fathers (Humphreys and Campo 2017; Mandel 2014; Stanley and Humphreys 2014). However, working directly with men who use violence and engaging them in behavioural change is not easy (Melchiorre and Vis 2013) and, in cases where the men are very dangerous, not always practicable (Juodis et al. 2014). There is a developing body of literature to guide engagement with fathers, which includes evaluations of perpetrator intervention programs and targeted fathering programs, such as Caring Dads (for example, McConnell et al. 2017; Panter-Brick et al. 2014; Stanley et al. 2012; Scott and King 2007). Engaging men through emphasising their roles as fathers can be a productive way to encourage them to reconsider their abusive behaviours towards both women and children (McConnell et al. 2017; Stanley et al. 2012), as fathers are more likely to take responsibility for the impact of their violence on children than on (ex)partners (Labarre et al. 2016). Additional benefits of engagement with fathers include greater awareness of the risks posed by the perpetrator, and possible shifts in fathering techniques, which may, in turn, influence child wellbeing (McConnell et al. 2017).

Working with Perpetrators of Domestic Violence in High-Risk Cases

In taking a whole of family approach, practitioners need to learn safe ways of engaging with men who may be dangerous, something that is often lacking in related higher education degrees (Humphreys and Campo 2017). In Australia, child protection services are incorporating the Safe & Together Model™ which provides training and coaching on practice (Healey et al. 2018). While it has been suggested that DFV perpetrators can be motivated to change their behaviours through the value they place on their relationship with their children, first they must understand the impact their violence has on children and parent-child relationships (Mandel 2014). Smith and Humphreys (2019) found that focusing on the relationship between men’s parenting and their use of violence is an important avenue for facilitating accountability; however, poor and inconsistent responses from child protection services can reduce the effectiveness of the approach. Deficiencies in system interventions can result in the system becoming the perceived problem (Heward-Belle et al. 2018), and perpetrators of violence failing to recognise the link between use of violence and statutory system involvement (Smith and Humphreys 2019). Additionally, perpetrator intervention systems often lack adequate power to compel DFV perpetrators to remain engaged, resulting in men exiting systems on their own accord in times of high risk (O’Leary and Young 2020).

Method

This paper reports the results from 15 participants (12 women and 3 men) who contributed to Community of Practice (CoP) workshops in Queensland, Australia, conducted as part of the Invisible Practices: Intervention with fathers who use violence project, funded by Australian National Research Organisation for Women’s Safety (ANROWS). The project, which ran from July 2017 to February 2018, aimed to investigate and simultaneously develop the workforce capacity of statutory child protection, men’s and women’s domestic violence services, non-statutory family support and justice services (police and probation and parole) in responding to fathers who perpetrate DFV. Research teams and participants were brought together for the project from four Australian states: Western Australia; Victoria; New South Wales; and Queensland. In total 65 practitioners and two consultants from the U.S. based Safe & Together Institute contributed to the research (Healey et al. 2018; Heward-Belle et al. 2019). All CoP participants in the larger project were senior practitioners, team leaders or managers.

The Safe & Together model centres on a perpetrator pattern-based approach, which includes:

  • Assessing the perpetrator’s behaviours and their impacts on children, including multiple and intersecting causes of harm to children;

  • Holding men to high-standards as parents; and

  • Focusing on men’s parenting choices (Mandel and Wright 2019, p. 121).

The three primary principles of the model are:

  • Pivot to the perpetrator;

  • Build an alliance with the victim/survivor; and

  • Focus on children (Humphreys et al. 2019).

The model contends that by focusing on perpetrator responsibility, workers and agencies can move beyond a mode of practice that blames victim/survivors for failing to protect children, ultimately increasing the number of children who can remain with the adult victim/survivor (Mandel and Wright 2019). Further, Mandel and Wright (2019) assert that a perpetrator-pattern based approach can increase worker safety, improve victim/survivor engagement and encourage greater collaboration across agencies. The specific research questions guiding the project were:

  1. 1.

    What do practitioners require from their organisations and/or other organisations to support them in working with fathers who use violence?

  2. 2.

    What evidence is there that the capacity building of CoPs, supported by coaching and supervision by the Safe & Together Institute, provide increased experiences of safety and support for practitioners? (Healey et al. 2018, p 12).

Participants came together to create regional CoPs by participating in six two-hour workshops across a period of six months. Prior to the CoP, participants received online training and two full days of face-to-face training on the Safe & Together Model, which ensured a mutual language and shared understanding of DFV. The decision to focus in this article on the findings from the Queensland CoP’s data was based on the unique inclusion of justice services (police and probation and parole) in this CoP. These services work with larger numbers of higher risk perpetrators, which resulted in the Queensland CoP discussing assessment and intervention at the higher end of the risk continuum than in the other three CoPs.

Each CoP was advised by the respective research team to bring de-identified representative cases from their case load to the workshops; these cases formed the basis of discussions among participants, including with the Safe & Together consultant who provided coaching on best practice related to cases presented from participants caseloads. Although there was no shared definition of risk across agencies, the unique presence of police and probation and parole in the Queensland CoP resulted in participants bringing forth cases considered to have very real potential for serious physical harm or for lethality among child and adult victims/survivors. These were complex cases participants wanted support with from the CoP. At the conclusion of the final CoP workshop, participants contributed to focus groups of 45 to 90 min duration. The aim was to seek their reflections on skills development and organizational capacity building through participation in the project.

A table showing attendance by sector is provided below. At two workshops, two guests from one of the family support services attended (noted as ‘+2’ in Table 1). Police and probation and parole are grouped as justice services to protect participant identity, as one service was represented by an individual worker.

Table 1 Number of workshop attendees
Full size table

An action research framework was utilised in the project. The approach is a blended strategy for inquiry and development, putting research and learning into practice and action in relation to a particular area identified to be in need of improvement. Action research is an interactive, cyclical process that enables change while the research is being undertaken (Wicks et al. 2008). Challenges are addressed through an iterative cycle of reflection and review. The approach was chosen for being strongly practice orientated.

Data were collected by the researchers through ethnographic note taking to document the policy and practice issues discussed in the CoPs and the focus groups. Qualitative thematic analysis was undertaken combining inductive and deductive analysis techniques. Given that research was occurring across multiple sites, specific deductive codes were applied to the data to ensure findings were generalised across the sites through use of a pre-prepared template, which provided a theoretical and practical basis for multi-site research (Stake 2013 in Healey et al. 2018 p. 20). The template also provided a systematic approach to data collection across the research sites. The overall interpretation and synthesis of the ethnographic notes was undertaken by one of the Chief Investigators at the University of Melbourne, with constant cross-checking by relevant research team members (Healey et al. 2018 p. 20). Examples of deductive codes included: professional practices, institutional policies, and language used to describe perpetrator behaviours. In addition to ethnographic note-taking, discussions in the Queensland CoP and focus groups were recorded and transcribed verbatim. Three of the four members of the Queensland research team, including a Chief Investigator, had undertaken data analysis in the larger Invisible Practices study, and were therefore familiar with the data. Each transcript was reread several times by the aforementioned Chief Investigator prior to undertaking qualitative thematic analysis using the process outlined by Braun and Clarke (2006). Codes were identified in the data; that is, inductively, with no use of the pre-existing template. Examples of inductive codes included: threats/danger; safety planning; fear/anxiety; and information sharing/communication. Codes were grouped to form themes. For example, the code ‘safety planning’ contributed to the ‘Learning from and partnering with women’ theme. The themes were discussed with the research team to check and refine each theme prior to the final analysis. The themes correlated closely with those reported in the larger study; these were:

  • Key skills identified for working with fathers who use violence and control;

  • Key factors identified in partnering with women;

  • Key skills in ensuring a focus on children and young people; and

  • The role of organisations and practitioner capacity building (Healey et al. 2018).

The language used in the findings section and discussion are gendered. This reflects the cases presented to the Queensland CoP, where all offenders were men and all adult survivors were women.

Findings

Four key themes in relation to risk assessment and safety planning in high-risk cases at the intersection of DFV and CP emerged from analysis of the data: learning from and partnering with women; applying a perpetrator pattern approach in high-risk cases; engaging with men as fathers; and the role of the judicial system in managing high risk cases. The action-research framework adopted in this study was seen to inform and support the integration of intra and inter-organization practice change, which required reflection and repeated application. Participants’ reflections on the process of change observed throughout the research (Wicks et al. 2008) is discussed prior to detailed reporting on the key emerging themes identified above.

Some participants were already involved with the same families prior to the project and therefore came with a history of collaboration. Participation in the project deepened and extended collaborative efforts with families deemed to be at high-risk, as illustrated by the following quote from a justice worker:

…prior to the training, we were all working collaboratively anyway. Just from having this training it has shifted the work. The information we share is a lot more detailed now. We are using information sharing a lot better. That’s been the biggest shift. We are going to more regular meetings, aren’t we now, focusing on our more high risk families.

A second justice worker’s comments confirm the value of extending training with CoP workshops:

Having that joint training together set the foundation…Then the workshops kept reinforcing the importance and, as we gained case examples and stories, reflecting back on the first one, everyone was very individual in their stories, to the point now where half the room actually know the family.

The workshops not only reinforced the value of collaboration, they helped participants integrate change into their regular practice, thus extending the training.

The workshops helped…you get so caught up with what you are doing, learning new things is harder…You can fall away with other urgent matters, but having the workshop every month prompted us. We were always really passionate about it, but having the workshops helped us incorporate it into part of our practice. (DFV worker)

The monthly CoPs allowed participants time to reflect and conduct their work differently prior to returning to a shared space for further dialogue, a process seen to improve assessment and management of high-risk cases. As noted first by a justice officer, followed by a quote from a child protection worker:

Having it monthly meant we really had to give it a go because we had to come back and talk about it as a group. I think it has all really made us come together, and that before families weren’t at the forefront when they really should have been. When we actually do meet up… we are sharing information better and we can see that those families are actually really high risk.

We have ongoing support. That’s the thing that has made it a learning tool and not just training, that ongoing support…this process is more useful than training (child protection worker)

Importantly, the action-research approach, in which participants reflected on practice as they implemented new strategies, was seen to improve outcomes for families. From a justice worker:

We had 50 high risk cases, and now we can see they are dropping off….So, we know it works because they are being moved from the high risk list to the other list.

Learning from and Partnering with Women

In order to effectively intervene and ‘manage’ high-risk cases, participants considered it imperative to partner with women, a key tenet of the Safe & Together model. Workers from child protection, in particular, drew attention to the assumptions they previously held about women’s actions, and how these underpinned punitive professional decision-making, notably the removal of children from their mother’s care. These assumptions and ensuing practices were contrasted with their current DFV-informed understanding of the strategies women living in high-risk situations employ to keep themselves and their children safe. This shift in practice was clearly illustrated in relation to mothers who allow a very violent man to enter or return to the home. Child protection workers drew attention to a phrase commonly used in practice, “She allowed him back in,” and argued that, without an understanding of the role of coercive control, women are held responsible for the continuation of violence, and for failing to protect their children, while focus on the perpetrator’s actions is diminished or rendered invisible. Changing practices involved workers dropping the ‘expert’ stance and learning from women. In some instances, this involved understanding that the mother’s choice to “allow him back in” may be the safest option for women and children living with a high, or potentially lethal, level of risk. A child protection worker stated:

Over a year ago, you wouldn’t get this work where you really listen to what the mum was saying.

Acknowledgement of women’s safety planning was an important step in partnering with victim/survivors and in managing risk. A child protection worker gave an example of the types of statements used in practice with women to demonstrate this approach:

We’re sure you were doing safety planning before we even came to your home. What kind of thing were you doing?

In understanding and respecting women’s safety planning, child protection workers empathized with women, built trust and rapport, and adopted some of the same strategies women used to ensure their safety, and that of their children, to manage their own anxieties and safety:

What we learned from her, I don’t think we’ll ever learn from anyone else again. She was amazing in what she taught us…we were really guided by her and listening. I think that we probably had a little bit of fear with engaging this man, too; you know what I mean. So, we were kind of walking along in mum’s footsteps and saying, “yeah”.

Participants described scenarios where women were previously viewed as not having appropriate boundaries and were judged for ‘over-sharing’ information with children. Such actions were reframed to show how women kept children informed and included them in safety planning:

She says that she wants the children to know because she wants them to be aware if he does come to the school. She’s setting up (identifying) risk factors for them. She’s actually setting up, “If this happens, then this is what you need to do.” She is looking after the children.

While participants recognised the safety planning women were undertaking, little attention was given to strategies children and young people employ or how they perceive or feel about the approaches others adopt on their behalf. It was noted that the removal of children as a means of protecting them from harm leaves their mothers without the monitoring role or support child protection workers can provide.

The children were removed out of the situation so it was actually a horrible situation because we then left her with him.

Applying a Perpetrator-Pattern Based Approach in High-Risk Cases

CoP participants regarded the perpetrator pattern-based approach to intervention in DFV cases, another core tenet of the Safe & Together model, undeniably beneficial to their practice. However, they considered it more complicated in high-risk cases. Mapping the perpetrator’s pattern of behaviour provides a systematic way of noting, recording and sharing specific and detailed information about the range of a perpetrators’ behaviours and tactics, which was important in the referral process of high-risk cases. CoP participants noted that in the absence of perpetrator mapping, high-risk cases can be closed by the statutory child protection service and referred to family support programs without appropriate information being conveyed, potentially leaving workers in these programs ill-informed and ill-equipped to effectively intervene and manage their own or the victim/survivors safety. A family support worker commented:

[Child protection] is closed; now you refer the family to other support services. We’re now going out to the home and trying to hold him accountable without necessarily all the background knowledge about everything related to that case, or necessarily a worker that is skilled in that.

Services systematically documenting and sharing detailed information about the perpetrator’s pattern of behaviour with police was stressed, as the following comment by a child protection worker illustrates:

They might have loads of information on him, but we just don’t know. It’s actually just leaving it up to chance that they know his behaviours and tactics, you know, and that’s really concerning.

Services generally develop and maintain their own documentation. A justice officer discussed an emerging plan to build on the perpetrator mapping approach through the use of a single document to be shared among service providers.

We do a lot of stakeholder meetings. Different stuff comes up every time with the high-risk cases. The action plan is to move towards making one document for these cases that we work on that we can all add information into. We will all be able to use the same document, where perpetrator behaviours are demonstrated, and incorporate mapping.

While the application of a perpetrator pattern-based approach was considered essential for determining and sharing understanding of the level of risk for women and children, the emergent pattern could exacerbate worker anxiety:

If you do the perpetrator mapping really well, if you just go and do it, all it does is draw your attention to just how unsafe and dangerous he is...Wow, now I’m really worried. I’m even more worried than before.

Similarly, from another worker:

…what we did for 3 hours, we sat in our office and we picked (worker’s) brain, and we wrote everything out up on the board, and that’s when we all sat there going, “This is a bit scary”.

Participants recognized that applying a perpetrator pattern-based approach in high-risk cases can reveal details about the perpetrator’s antisocial and criminal behaviours, and that this information can be difficult to use in practice, including in the legal arena, due to concerns for the safety of victim/survivors.

When she got four pages of his behaviours she was then thinking, “Oh my God, we can’t actually use this information”. How can we use it because if he gets hold of this information she’s really unsafe? Now what does that mean for taking orders and evidence and documentation?

Participants reported that women living in high-risk situations will withhold information about perpetrator’s criminal and other behaviours in order to protect themselves and others, including workers, indicating that potentially important information on perpetrator behaviours remains hidden:

She had a lot of information about his offending behaviour. She knows how he hides guns, she knows a lot of stuff that she will not even talk to us about; she won’t talk to anybody about, because it’s so dangerous.

Engaging with Men as Fathers

While the men referred to in this paper were regarded as high risk, they posed various levels of threat and differed in the extent of their participation in criminal and/or other antisocial behaviours. The Safe & Together model acknowledges differences among those who perpetrate violence and uses the concept of ‘pivoting to the perpetrator’ to encourage professionals to differentiate between perpetrators, and to target their interventions in accordance with each individual perpetrator’s pattern of behaviour. While pivoting to the perpetrator would, in many cases, result in conversations with men about the impact of their violence on family functioning, workers in the CoP drew attention to the limitations of a blanket approach to direct engagement, deeming some men too dangerous. A child protection worker recounted a conversation in which a mother stated that workers attempting to engage her partner would exacerbate the risk she lived with:

She said, “No, you should do nothing right now. If you say anything, or do try and engage him - that is going to make me more unsafe”.

Participants reported that using fathering as a motivator for change, an approach consistent with the Safe & Together model, is less effective among high-risk cases. Workers recognized their disinclination to use the approach with men who show no concern for their children, no remorse for harm they caused, or any interest in moderating their behaviour. Such men were considered to be particularly dangerous:

My challenge is that the clients I work with, their partners have committed violence at the highest end of the scale, and it has almost gone past how much we care about their fathering; it’s more about payback and revenge. You’ve gone past the point of getting them to change, when many of them just don’t care. The kids are collateral.

Similarly from another participant:

Because from our perspective we are always dealing with the high-risk families, and they have the attitude like, “I don’t care. I don’t care what I do wrong.”

Workers needed to make decisions about which men were ‘workable’ and those who were not. Among the former, participants commented on developing the ability to engage with men who use violence, but noted that these abilities were still in their infancy:

It is such a new skill, it’s finding the repetition and mentoring and reflection, because it is scary… For me it is practicing the actual skill. And if you stuff it up, then he is gone. So if you say the wrong thing, you know it’s over. So you are practicing live, and you don’t want to ruin the opportunity. How to say things. (Women’s DFV worker)

Among the latter, holding men accountable through direct engagement was difficult to achieve. Men who deny their behaviours were considered “the hardest dads to work with”, but beyond these men were those whom workers did not even attempt to engage due to the perceived level of danger to women, children and workers:

We haven’t done any engagement with him due to her safety. And it’s almost like we can get stuck in that place. So we can get mum out, kids, do that great work with the DV service, and then, right, what are we doing about him?

Understandably, dangerous men generated fears for workers’ own safety and resulted in practice that diverted attention from the ability to monitor the perpetrator, let alone engage with him, leaving these men beyond the scope of intervention:

In those really high [risk] cases, people get so nervous, it’s like everybody’s covering their own arses and not actually working to what they need to do to actually keep an eye on him and address him.

The Role of the Judicial System in Managing High-Risk Cases

While participants acknowledged the need to monitor and manage high risk cases, they reported this is compromised by inadequate or delayed communication and collaboration, particularly between justice services and the courts. While attempts to improve practice were noted and appreciated, the sheer volume of cases compromised the safety of women and children. A justice officer reported:

They’ve got so many cases going through every day, even the high-risk ones are sort of slipping through where they’re not even communicating back to put red flags on them people. And sometimes the domestic violence orders are taking two to three days to actually receive them and put them on, so it’s really hard to monitor when we’re not getting that information back from the court.

The courts’ lack of understanding of the dynamics in DFV violence, particularly the use of coercive control, was considered problematic. While CoP participants reported formerly approaching their work without a domestic violence informed lens and holding women responsible for the continuation of violence, as mentioned above, they argued that the courts still fail to recognize or understand the protective measures women take in response to perpetrator behaviour, noting how these actions can be misunderstood and misconstrued. A justice officer stated:

I can understand why she’s doing that (writing to the incarcerated perpetrator in order to monitor his reactions and to appease him), but the court, from a defence, from his barrister, who is going to say, “Hold on a minute, she’s leading my client on”…He might get off on this because of that defence. I can see why she’s doing it, from a safety perspective. He’s very dangerous.

The quote below from a child protection worker illustrates how women and children at high risk of harm can fall between the gaps when magistrates are ill-informed, do not have a DFV informed lens through which to understand the tactics of coercive control, and/or families have simultaneous involvement with several courts that do not communicate:

We had some matters before the Children’s Court but we know that the matter, the domestic violence, was before the Magistrates’ Court…we had to remove the children because one magistrate amended the domestic violence order to allow dad to have contact, upon mum’s written consent, and we absolutely knew the coercion behind that…We also knew from our work with our partners that there were Supreme Court bail conditions that this father was not to have any contact with any of the children because of an incident that happened earlier on in the year…they (police prosecutors at Magistrates’ Court) had no idea there were Children’s Court matters and there were actually Supreme Court bail conditions. They were simply going in to hear the DVO (domestic violence order). So we’ve really had to explain this quite quickly, “No, no, no! You really need to be aware there’s all these other things pending, and he’s a very dangerous man.”

While child protection workers often have information that could be used to support police prosecutors, CoP participants reported that threats to worker safety led to reluctance to provide information in high-risk cases, which compromised intervention efforts:

Child protection have got more information than police…for some of the more high risk cases we’ve been saying, “Well, we would prefer you guys to put it in (the report) because you can, obviously, provide all the evidence to actually support this order”…Child protection have been reluctant to do that…we’ve got high risk perpetrators who would single out the child protection worker…they’d single out those workers and put them in an unsafe situation.

Participants pointed out that in high-risk cases, justice efforts can be inadequate in protecting women. In such instances, it was argued that women live in entrapment as a means of managing their safety:

Mum in this case is really smart because she has worked out that no one, police etc., has been able to protect her. So, I think, it’s part of her safety plan, to learn to live with him to make her safer.

Of concern was the threat dangerous men could pose to women, children and workers, even when incarcerated. A justice offer described the continuous harassment of a mother by her incarcerated (former) partner:

Over the last month, he has rung her 300 times from prison…between 7 o’ clock in the morning and 4.30 in the afternoon, he is literally calling her, and calling her, and calling her, and calling her.

The lack of communication and collaboration between the prison and workers in relation to safety planning for women and children was also evident. A child protection worker’s comments highlight this significant gap in practice:

Dad is trying very hard to threaten her from prison…he’s just sitting there, you know, seething, and, so, for her, I do think, like, what’s our communication with prison about what’s going on in there for him? There seems to be a massive gap. There’s no communication with us and prison, what programs he would do, is there going to be an assessment on him before he gets out in regards to how he’s feeling about her and the child?

Discussion

The present study is one of the few accounts of the process of safety planning in DFV and child protection cases (Logan and Walker 2018). Action research methodology was adopted in acknowledgement that research can learn from practice-based knowledge, and help to inform further practice developments. The findings distil key learnings for practice at the intersection of high risk domestic violence and child protection concerns. It is evident that practitioners valued partnering with women, allowing victim/survivors to inform safety responses. In doing so, practitioners needed to work around organisational policy and practices that can result in oppressive interventions, such as directing women to separate from dangerous men, or removing children. Partnering with women reduced the tendency to engage in practices that “mirror the tactics of power and control exerted by perpetrators” (Heward-Belle et al. 2018, p. 145). A shift away from oppressive policies and practices will require the development of domestic violence proficient practice within the range of organisations charged with ensuring the safety and wellbeing of women and children. This translates into services managing risk while simultaneously holding perpetrators responsible for their own behaviour and the safety of their children, rather than blaming victims/survivors (Heward-Belle et al. 2018; Mandel 2014), and placing unjust expectations on them (Wilson et al. 2015). These shifts in practice need to be mirrored in organisational protocols to allow statutory and other workers the freedom to be guided by victim/survivors strategies when safety planning (Logan and Walker 2018), even if these actions seem counter-intuitive.

Partnership with victim/survivors can make engagement with men using violence possible, as the pattern of behaviour can be made visible and risk management strategies can be considered (Humphreys et al. 2019). Yet, research has identified that, in their day-to-day experiences of safety planning, workers can be judgemental about victim/survivors decision-making (Logan and Walker 2018). Similarly, practitioners in the present study identified pitfalls in safety planning: being judgemental towards the victim/survivors decisions and dictating to the victim/survivor rather than working collaboratively. As Heward-Belle et al. (2018) note, services can pathologize victim/survivors responses to abuse and oppression, including the strategies they use to keep themselves, and their children, safe. Practitioners in the present study privileged women’s perspectives (Jenney et al. 2014; Stanley and Humphreys 2014), regarding this the key to successful intervention in high-risk cases. However, Wilson et al. (2015) caution that DFV responses based on an empowerment approach that privileges victim decision-making can deprive women and children of appropriate support, particularly in high-risk cases of potentially serious or lethal harm. Campbell’s (2004) work highlights that approximately half of women who were killed or almost killed by an (ex)partner accurately predicted the risk to themselves. In using women’s judgements, in tandem with information from partners, practitioners made informed decisions about whether to attempt engagement with individual men and how to go about devising safety plans.

The findings of this study contribute to the developing body of literature on engaging fathers who use violence, adding the voices of practitioners who do not work directly in men’s behaviour change or fathering accountability programs to those who do; for example, practitioners delivering the Caring Dads program in Canada, Australia, the United States, and Europe (McConnell et al. 2017). While the latter group of practitioners can draw on the motivators for engagement used in these programs, workers in the wider service system can face different challenges when engaging fathers. They can, nevertheless, contribute to the safety of women and children by working alongside those directly engaged with fathers, with other agencies and with families. This is particularly important when men do not make sufficient behavioural changes, or change only temporarily (McConnell et al. 2017 p. 413). Perpetrator mapping could provide a platform for shared conversations among professionals, as indicated in the present research. It could also be used with family members to garner the full range of behavioural impacts on child and adult victims.

Deciding whether to directly engage perpetrators as fathers requires specific knowledge and skills that are not routinely included as core learning in training qualifications for human services and social work (Humphreys et al. 2019). Going beyond practice norms could lead practitioners to engage in what Heward-Belle et al. (2019) refer to as dangerous practices. This term is used to cover the risks in engaging with men using violence, and the dangers to workers who go beyond their organisational remit when engaging men in high-risk cases (Heward-Belle et al. 2019). The findings in the present study indicate limits to the strategy of using men’s fathering as motivation for change. Among the men in this study were some described as “seething” and more concerned with “payback” and revenge” than parenting. Such proprietary revenge is a red flag for potential child and/or maternal homicide (Juodis et al. 2014). In such cases, even monitoring was barely possible, confirming Melchoirre and Vis’ (2012) assertion that it is not practicable to work with all men. Perpetrator mapping could be usefully applied in identifying which men have behaviour indicative of psychopathic traits (Juodis et al. 2014) and, if conducted collaboratively, could lead to a shared professional understanding of the level of risk these men pose for women, children and practitioners. Olver and Jung (2017) note the benefits of using more than one risk assessment tool. Perpetrator mapping, with its emphasis on assessment of the perpetrator’s behaviour on family functioning, more broadly, could augment the use of validated risk assessment instruments by extending assessment to include impact on children’s wellbeing. In this way, it could increase inter-agency collaboration and help to ensure children receive the services they need, as recommended by Reif and Jaffe (2019) in their research on domestic violence fatality review teams.

Practitioners rely on collaboration and the information it provides to assess risk not only to family members, but also to themselves (O'Leary et al. 2018) and to enhance the efficacy of perpetrator interventions. This requires accurate information from all available sources; however, agencies differ in their sources of information, how they use information, and their sharing of it (Juodis et al. 2014; Stanley and Humphreys 2014). The findings from the present study clearly demonstrate the need for what Juodis et al. (2014) refer to as ‘collateral information’ (p. 385), which may not be readily available through administration of a risk assessment instrument alone. However, caution needs to be exercised. CP services, in particular, were often privy to details about perpetrators, including criminal activity and other antisocial behaviours, but disclosing information could jeopardise the safety of individual workers. In such instances, police may need to gather evidence from other sources and include it in their own statements in applications for protection orders. To advance collaboration, participants recommended that perpetrator mapping be undertaken using the one key document with input from justice, child protection and domestic violence services. This recommendation has the potential to improve the safety of adult and child victim survivors and their associated workers, as information could be put forward without, potentially, identifying the source. A shared approach, with detailed discussion and planning for women and children’s safety, as well as their own, may also help workers overcome the anxiety that perpetrator mapping can generate when the full history and range of abusive and dangerous perpetrator behaviours are collated. This strategy could make a valuable contribution to the functioning of coordinated community responses critical in preventing lethal and non-lethal DFV (Juodis et al. 2014). Having clear understandings of legislation and policy on information sharing would be prudent.

Managing high-risk cases across agencies and holding men who use violence accountable for their actions requires frequent contact with perpetrators and victim/survivors by health, social service or criminal justice professionals (Juodis et al. 2014); however, the findings from the present study highlight gaps in systemic responses including statutory child protection services closing after families are referred to non-statutory family services without the conveying of sufficient information. Similar results have emerged from research on perpetrator intervention systems quickly losing track of high risk perpetrators leaving women in the invidious position of needing to know his whereabouts to manage their safety (O’Leary and Young 2020). While accounts from CoP participants indicate that Australian services are continuing to generate solutions to challenges in practice at the interface of DFV, child protection and policing (Stanley and Humphreys 2014 p. 83), it is also evident that the legal system is obstructing the seamless or wrap around support women and children need and thus “failing to curtail the perpetrator’s ability to be abusive” (Wilson et al. 2015 p. 26). The Australian legal system is reported to be patriarchal and to mirror the tactics of coercive control deployed by abusive men. Further, judges are said to not understand the dynamics or the impact of domestic violence on women and children (Heward-Belle et al. 2018, p. 142). A fragmented legal system is evident in the present study. Different legal and service jurisdictions often have information about the same family that could provide a more comprehensive risk assessment, but often this information is not meaningfully shared. Stanley et al. (2011) note that the information needed for coordinated responses to the needs of women and children affected by DFV “is not easily accessed and has to be mined from a range of sources”, and that it is necessary to “dig deep for information” (p. 2387). The findings from the present study show that child protection workers often held this information, which they obtained by partnering with women, but that disclosing details to police could jeopardise their personal safety, and that of women and children. Until the various courts adult and child victim/survivors, and perpetrators, may be involved with become domestic-violence proficient, and intervention is undertaken collaboratively with due consideration of the risks for individual workers, particularly when providing information necessary for law enforcement to effectively deal with abusive men, integrated responses will be hampered in their efforts to protect vulnerable women and children.

It is noteworthy that children and young people were largely absent from the discussion of high-risk cases presented in the CoPs, even by practitioners employed at child-focused agencies. Practitioners did not include details of discussions with children, instead working on the assumption that partnering with adult victim/survivors would always include the perspectives of the child. This does not account for instances where the needs of the young person and the protective parent diverge. Further, the impact on children may not be fully understood as it may not directly reflect the severity of an incident (Stanley et al. 2011). Horton et al.’s (2014) research with practitioners on their perceptions and experiences of safety planning with children exposed to DFV demonstrated limited understanding of age appropriate strategies for working with children, not knowing how to safety plan with children, or how to include parents as well as children. However, positive trends have been noted amid concerns that child-inclusive practice will place too large an onus on children to be responsible for their own safety (Horton et al. 2014), a concern that would have particular significance in high-risk cases. Reif and Jaffe (2019) call for future research and practice into domestic violence homicides to keep children visible in preventive strategies and responses. Future research could build on Scott et al.’s (2020) work into child homicide in the context of DFV, which calls for standardised lethality risk tools to be used in child-focussed agencies.

Limitations

The present research goes some way towards addressing gaps in the literature on intervention in high risk cases by documenting examples from practice. The need for further consideration of how cases are defined as high-risk in domestic violence cases is highlighted. While there are working definitions used by practitioners on the ground, the criteria have not been documented or critiqued in the academic literature. A review of the protocols of participating agencies on what constitutes a high-risk case could have strengthened the paper, and ensured there was a consistent definition of high-risk applied to the examined cases; however, CoP provided a process for checking the validity of practitioner insights across multiple systems. The effectiveness of the outlined strategies is unknown: workers’ accounts suggest successful partnering with women, but the potentially corroborating, or disconfirming, voices of women and children were not included in the study. The CoPs were designed to increase workforce capacity and create shared understandings of key concepts, such as partnering with women, across agencies. These key concepts were delivered by coaches trained in the Safe & Together model; however, a critique of the model and evidence of its independent evaluation have not been included in this paper.

Conclusion

This paper adds to the evidence on the strengths and limitations of the Safe & Together model in practice. It also highlights avenues for further development of service sector responses in high risk cases, and identifies future research directions. This paper has argued that practice needs to be informed by the victim/survivor perspective, which requires collaboration with women, not coercion. The findings suggest workers are safer when they incorporate women’s voices, which is only possible through partnering with women and fully understanding the daily steps they take to keep their children and themselves safe. The findings confirm that relationships with children are not a motiving factor for behaviour change among the most dangerous men. Also evident is that lack of communication and collaboration between the various service sectors and the courts, and between courts, is hampering wrap-around support for women and children, and increasing risks to worker safety. Training in the Safe & Together model for all judiciary is likely to promote a more informed and integrated system for managing high risk cases. Children were notably absent in practice discussions on high-risk cases. Further research is needed on child participation in inter-agency responses to DFV, especially in complex situations where the needs of the adult victim/survivor may differ from those of the child or young person (Stanley and Humphreys 2014). Finally, this paper adds to the growing body of literature on improving practice when engaging men using violence and makes recommendations for advancing collaborative practices.