Baseline characteristics and outcomes of the main perpetrator programme within the Hampshire Domestic Abuse Prevention Partnership, UK: A mixed methods study

Effective perpetrator programmes should be embedded within a community response, engage all types of perpetrators and involve an educational approach that integrates the survivor’s voice. The Domestic Abuse Prevention Partnership (DAPP) is a transformative partnership based in the UK that aims to provide an integrated approach for perpetrators and survivors. This pragmatic mixed methods study was conducted to examine the baseline characteristics and individual outcomes of the main perpetrator programme within the DAPP. Applying a triangulation design, routine police re-offending aggregated data, pre- and post- perpetrator programme questionnaires, in-depth interviews with survivors, and focus-group discussions with perpetrators (clients) were integrated. Statistical analysis and thematic analysis were applied to quantitative and qualitative data, respectively. The majority of clients (47%) referred through the DAPP (n = 228) described wanting to make their relationship better as the main reason for engaging with the main perpetrators programme. Post-perpetrator programme questionnaires identified positive changes in both emotional behaviours and physical behaviours amongst clients, which were also supported by examples of improved relationships with their children described in survivor interviews. Three themes were described: first, making positive progress; second, impact of the children’s module; and concerns around sustaining new behaviours. Over the monitoring period, 1 in 5 clients were either suspected or convicted of domestic abuse crimes following the programme. This suggests that further maintenance of positive behaviours and reinforcements are required for some clients. Given that clients felt children were a strong motivating factor for completing a programme, it seemed paradoxical that no specialist services were made available for them. Future reiterations of the DAPP model should at least address how best to work with children in families where domestic abuse occurs.

Introduction National statistics for England and Wales estimate that 6.0% of adults aged 16-59 experienced domestic abuse in the last year, with a higher proportion experienced by women (7.9%) versus men (4.2%) [1]. These estimates are based on the latest UK government's definition of domestic abuse, as 'any incident or pattern of incidents of controlling, coercive, threatening behaviour, violence or abuse, between those aged 16 or over who are, or have been, intimate partners or family members regardless of gender or sexuality' [2]. Domestic abuse has wider cost implications for the health and social care system, with direct medical and mental healthcare costs approximating £1730 million per annum in the UK [3,4]. Intimate partner violence is particularly prevalent among both survivors and perpetrators receiving mental health services. Although the relationship is likely bi-directional, domestic abuse is associated with mental health problems, including depression and post-traumatic stress disorder (PTSD) [3,5,6].
Given that the majority of domestic abuse survivors are women, the development and implementation of perpetrator programmes often take a gendered view by examining the role, and influence, of power and control within the relationship. Most domestic violence perpetrator programmes (DVPPs) are therefore embedded in an understanding of heterosexual relationships; apply cognitive behavioural, pro-feminist or psychodynamic treatments; whilst 50% reportedly use a combination of multiple treatment types [7]. Over the years, DVPPs have developed through a process of critical reflection and engagement using the National Practitioners Network (NPN), as well as from listening and responding to perpetrators [8].
Voluntary community based programmes work with perpetrators of domestic abuse referred from community services, such as Children's Services and the family courts. Such programmes often involve multiple agencies, meaning that there is increased opportunity for inefficiency, duplication and poor decision making [9]. Many community agencies, with a focus on survivors, also lack the adequate knowledge and competency of working with perpetrators. To circumvent some of the issues around dealing with perpetrators, voluntary community based programmes may use co-location, whereby referral agencies are supported through additional training and engagement. Given the complexity of embedding DVPPs within the community, and concerns around quality and efficiency, there is particular interest in modelling a multi-agency approach and evaluating the process of undertaking DVPPs, in addition to reporting the outcomes of such programmes [10].
The effectiveness of perpetrator programmes has long been debated. However, it is difficult to estimate the effectiveness of these programmes, primarily due to the difficulties arising from study methodology. Limitations in study design include small sample sizes, low response rates during follow-up, and no equivalent comparison or control group [11]. However, evidence suggests that programmes embedded within a coordinated community response, that aim to identify, treat and retain perpetrators, lead to the most positive response in terms of re-assault prevention [12].

The Domestic Abuse Prevention Partnership (DAPP)
The DAPP is a multi-agency initiative based in Hampshire, UK led by a third sector organisation. The DAPP is a novel community based partnership that provides a tiered, flexible and needs-led approach; bridging specialist services for domestic abuse, with a focus on working directly with perpetrators. The DAPP is considered novel as it provides a voluntary programme (i.e. not court-mandated) for perpetrators and ensures that survivors are offered support at the same time. Furthermore, the DAPP provides one to one support for perpetrators whose lives are considered too chaotic to manage group work; in addition to the psycho-educational support offered through the one to one work, such individuals may be given additional support to find housing or employment, for example.
Under the DAPP, perpetrators (clients) engage with a 20-week programme, named the RADAR/ADAPT programme, consisting of modular group work. The modules include awareness raising, relationship dynamics, children and domestic abuse and parenting.
The aim of this study was to add to the limited evidence around community perpetrator programmes, and what they achieve, by evaluating motivations for engagement, type of abuse committed, as well as behavioural change outcomes following the 20-week programme, using data from both client and survivor.

Methods
This was a mixed methods study with a triangulation design model, which aims to examine the effectiveness of the DAPP model by simultaneously triangulating both quantitative and qualitative data [13].

Police data on reoffending (Record Management System)
Police reoffending data was used as a proxy indicator of harm caused to the partners, or former partners, by those who had attended the main perpetrator programme within the DAPP. Hampshire Constabulary provided aggregated police monitoring data on 80 clients who completed the programme between 01 April 2016 and 30 September 2017. The data included the total number of clients linked to any crime, as well as domestic-abuse related crime, either suspected or convicted.

IMPACT monitoring (Pre-and post-programme questionnaire)
The IMPACT monitoring toolkit was developed by the Work with Perpetrators European Network to harmonise and enhance the monitoring and evaluation of work with perpetrators across Europe [14]. The IMPACT toolkit consists of a questionnaire that examines, through self-report, perpetrator behaviour (towards their partner or former partner) including type of abuse; their well-being; the impact on children; and their current relationship with their partner. To monitor progress through the main perpetrator programme offered through the DAPP, clients filled out the IMPACT questionnaire before (T0), and after, the main programme (T3).

In-depth interviews
The qualitative methods included focus group discussions and in-depth interviews. The indepth interview was chosen to explore outcomes for survivors, whose partner (or former partner) had engaged with the main programme. One-to-one interviews were considered the most appropriate means for exploring personal experiences and perceptions, as compared to any other qualitative methods. A support worker from the third sector organisation contacted and identified survivors who were willing to be interviewed. The majority of survivors did not have regular contact with their perpetrator and therefore could not provide accounts of the programme. Based on the difficulties of consenting survivors to be interviewed about their partner's (or former partners) engagement with the programme, and the limitations in numbers, a maximum of 10 participants were to be recruited. The inclusion criteria included: survivors that acknowledged incidence of abuse through case reporting; survivors that were currently in contact with staff at the third sector organisation; and survivors that expressed willingness for feedback regarding clients' progress through the programme.

Focus group discussions
Focus group discussions were considered as the most appropriate way to draw on the experiences and perceptions of the majority of clients-as their experiences of the RADAR/ADAPT programme were constructed as a group. Due to the nature of running a 20 week programme, it was not possible to choose groups at random, as selection occurred during a short research period window of three months, when only a small number of groups were running. Based on feasibility, a maximum of two focus groups were selected, and took place before, or after, the planned programme session.

Interview protocols
All in-depth interviews were carried out by SM, and the focus group discussions by both SM and BM. Whilst all participants engaged within focus group discussions were male, it is also necessary to acknowledge that both SM and BM are female researchers. A semi-structured topic guide was developed based on the literature, with both structure and flexibility, so that the order in which the topics were discussed followed a natural course. All interviews took place in a private location without any outside disturbances. Consent was required from each participant before the start of the interview. The recorded interviews were later transcribed.

Data analysis
For quantitative data, statistical analysis was undertaken using Stata version. 14. For both IMPACT and police data, which were based on individual data, complete case analysis was used. Using the IMPACT data, descriptive statistics were used to summarise the client's individual variables. Clients were asked about their behaviours reporting on a Likert scale: never, sometimes or often. There were a total of 14 descriptive questions of physical behaviour; 11 of emotional behaviour; and 7 of sexual behaviour. These responses were recoded into binary codes: 0 -never; and 1 to include those that responded either sometimes or often. These were then used to produce an overall continuous score for each behavioural type (emotional, physical, sexual) and are presented as prevalence (%).
To examine change over time, individuals were matched at T0 and T3, and behaviours were compared. Reoffending rates were determined using aggregated police data for those completing the programme over one year; showing the total crimes committed (as suspect or convicted) amongst completers, the proportion of total crimes that related specifically to domestic abuse, and the total number of clients (completers) that were implicated.
For qualitative data, the transcripts were exported into NVivo version 10, a qualitative data management software package. Using the thematic approach outlined by Braun and Clark, a six phase process was undertaken; in which patterns were identified, analysed and reported within the dataset [15]. Furthermore, negative case analysis was undertaken to look for patterns that contradicted patterns in the dataset.

Research ethics
Ethical permissions were granted by the University of Southampton Ethical Committee (ID: 26211). Upholding ethics, good governance and quality in practice was central to all processes. These include the moral principles guided by four main principles of bioethics-as suggested by Beauchamp and Childress [16]. The principles are autonomy (informed consent), nonmaleficence (do no harm), positive beneficence (benefits of research outweigh the risks) and justice (research strategies and procedures are just and fair). The social workers consented for routine data, from both clients and survivors, to be shared for the purposes of the evaluation by using a consent tick box on an assessment form. This routine data was shared for the research purposes as wholly anonymised data. During the evaluation, analysis, and write-up, no data was linked to the names of participants. For qualitative interviews, formal informed consent was obtained by researchers (SM, BM) for each participant. For survivors, the research interviews were conducted by phone for ease and privacy; verbal consent was obtained and audio-recorded. For clients, written consent forms were obtained. The qualitative quotes have been pseudonymised in this written report. All interview materials were approved by the Ethical Committee.

IMPACT questionnaire
Between April 2016 and November 2017, 228 individuals completed the IMPACT questionnaire at baseline (T0) before the start of the programme. Over the same period 80 individuals completed the programme. The remaining 148 either failed to start or failed to complete the programme. The following Tables 1 and 2 show the reasons for coming to the programme at T0.
The highest proportion of clients were no longer with their partners (33.8%), whereas approximately 1 in 4 individuals were still in a relationship with their partners and living together.
As shown in Fig 1, 7% of clients exhibited none of the total eleven emotional behaviours listed. However, 24% exhibited only 1 behaviour and approximately 20% exhibited 4 or more of the total emotional behaviours.
As shown in Fig 2, 49% of clients exhibited none of the total 14 physical behaviours listed. By contrast approximately 25% of clients exhibited 1 physical behaviour of the total listed. As shown in Fig 3, approximately 89% of clients reported exhibiting none of the sexual behaviours out of the list of 7; and only 11% reported exhibiting one or more sexual behaviours.
A total of 34 clients completed the individual IMPACT questionnaire following completion of the RADAR/ADAPT programme. Fig 4 shows the difference in the number of emotional, physical and sexual behaviours before and after the programme. As shown in Fig 4A compared to before the programme, fewer clients had higher emotional scores, and the number of Table 2. Relationship status of those assessed before the start of the programme (self-report, using the Impact questionnaire (n = 228). individuals that exhibited two or less behaviours increased. Fig 4B shows the difference in the number of physical behaviours before and after the programme. After the programme, fewer individuals had higher physical scores, and the number of individuals that exhibited 0 behaviours increased. Fig 4C shows the difference in the number of sexual behaviours before and after the programme. Responses to sexual behaviour appeared to be very similar before and after the programme. Police data. Summary statistics are shown in Table 3. Over the monitoring period covered, approximately 28% of clients were linked to crimes (as suspect or charged) after completion. However domestic abuse amounted to less than half of these total crimes (43.2%). Of those who completed, 17.5% of clients were linked to domestic abuse crimes after completion.

RELATIONSHIP STATUS Frequency, n Prevalence, %
Qualitative evaluation. Overall two focus groups (n = 12) were conducted with clients (perpetrators) who had either completed, or were close to completing, the programme. A further eight survivors were interviewed during data collection, to explore the extent and nature of any behavioural change observed in their partners, or former partners. Of the survivors interviewed, approximately five were in contact with the partner concerned. Overall there were three major themes, which are outlined in further detail below.
Theme 1-Making positive progress. A richer understanding into how the RADAR/ ADAPT programme brought about change was gained through focus group discussions. The clients were asked about how their behaviours had been influenced by the programme. Overall, clients indicated that they were making positive progress by attending programme, that the RADAR/ADAPT programme had a big impact compared to other treatments (sub-theme 1), and that both the effort and time spent on the programme were integral to such progress (subtheme 2). One respondent felt that the programme had enabled him to accept responsibility for his behaviour and suggested that the relationship with his wife had improved as a result of this, "Obviously I can't speak for my wife, truthfully, but I do touch wood that our relationship has started to actually repair, to at least make positive progress. Not least because I've actually accepted responsibility for what I was actually doing, and realising that my behaviour just was completely unacceptable. I think almost that has had-I'm trying to think, but that has been part of what the course has actually worked on me to actually come to terms with."

(Client 4, Focus Group Discussion 1)
Based on survivor feedback, the majority of respondents did not observe or experience any on-going abuse whilst their partners, or former partners, were attending the programme. By contrast, the majority of survivors observed some degree of behavioural change when discussing their relationship, The same respondent goes on to acknowledge that this change was initiated through an increased sense of self-awareness and understanding of where blame lies, "Yes, he's definitely, his attitude towards life, I think, has changed a lot more. He's actually realised that it was him, because I think that's a major step in that sort of self-awareness in domestic violence where they have to realise that it's them, because his before was like, 'Oh, it's not me, it's you, you make me do this, you're the one who made me angry.' He actually stops and thinks about things before he says them and his just whole perspective, the whole, I don't know, it's weird! Definitely a lot calmer though and he's able to talk a bit more now. . ..It weren't that long after, to be honest. It was probably about a month after the course." (Interview with survivor, 004) https://doi.org/10.1371/journal.pone.0218408.g004

Table 3. Summary statistics on crime associated with 57 individuals completing the RADAR/ADAPT programme over one year; number shown in brackets.
Percentage of total people who completed the DAPP programme linked to crime after completion, % 28. In relation to past behaviour, one survivor identified the power and control as an issue in their relationship. Since her partner had been attending the RADAR/ ADAPT programme, however, he had exhibited less controlling behaviours towards the family, "It was very much about power and control before. Very much he had to be in control of everything and he got very stressed if things weren't as he thought they should be or people didn't do or the children or me or I didn't do things the way he thought. . .."

(Interview with survivor, 002)
On explaining the impact of the change, one client felt that his wife was less fearful of their relationship, "She suddenly came out with the phrase, 'You've really changed.' Now, I couldn't see that. I think I've changed. I think, in other words, I could change, but she actually could see. To me that was really positive, because it has made her, because she was very fearful, and suddenly she wasn't fearful anymore. Suddenly she realised that I was nothing to be afraid of, anymore. 'He's changed enough for me to be able to relax and trust,' which was great, really."

(Client 4, Focus Group Discussion 1)
However, one respondent felt that despite the number of apologies offered by her former partner, certain behavioural traits still persisted, such as anger,

"Still quite volatile and he has apologised a lot-which is something he's never done beforebut then he'll follow that with getting very angry if I don't do what he wants after he's apologised. He's still very manipulative with the children and with me. Yes, he is quite prone to getting very angry."
(Interview with survivor, 005) Subtheme 1.1 -The influence of the programme on individuals compared to other treatments. Another example of positive progress driven by this programme was given by one survivor who felt that, compared with the other types of therapy her partner had sought in the past, the RADAR/ADAPT programme had had the biggest influence on her partner's behaviour. This view was also corroborated by clients on the programme; a few described how influential the RADAR/ADAPT programme had been on their lives, compared to other therapies.

Subtheme 1.2 -Positive outcomes due to engagement with programme.
Further to the positive progress made to behaviour, clients' commitment to the programme also had an impact on their individual circumstances. One client suggested that the effort and time put into the programme was acknowledged by the judge reviewing his custody case. He further described how his relationship with both wife and children had improved, "Like yourself, I have great shame that I've affected and hurt my children. Coming from a place of seeing them supervised six times a year for a couple of hours, to seeing them now pretty much when I want, or when they want. I think it was the reports that [XX] and her colleagues filled in that I could present to social services, then they questioned me, and examined and interviewed me. They made a report, and these two reports go to the courts. Basically, the judge has got something to go on then, hasn't he? He said to me, 'You've done the work. You've put this effort in. We've got a report from the programme. We've got a report from social services that interviewed you,' and now it's a family on the mend. My wife and I are amicable, and I'm seeing the children on a, basically, when they want, really. I mean, they want to go fishing, they want to go swimming. They're two boys, ten and 12. I mean, they're enjoying it now, which is great, and they're coming to stay with me at Christmas."

(Client 2, Focus Group Discussion 1)
Another client explained that since starting the programme his ex-partner had contacted him, encouraged by his investment in the programme and the positive steps he was making, The reports from the main voluntary sector organisation and partner agencies also facilitated this process; relationships seemed to change, or improve, as a result of attending the programmes and having reports shared with social services, ". . ..but because social services are involved, they actually presented the reports to my wife, and made their own report. If not, no one would have really explained what I'd been doing. Even then, even though the report was given to her and explained to her, it wasn't until about three months after court, and the relationship started to improve, and I was seeing the children normally, and my wife and I were amicable."

(Client 3, Focus Group Discussion 1)
As conveyed in the quote above a positive outcome, or a noticeable change in relationship, could take months.
Theme 2-Impact of the children's module. The children's module had a powerful influence on clients attending the programme, as represented in this theme. One survivor described the emotional influence of the RADAR/ADAPT programme on her partner, particularly in acknowledging his past behaviours towards their child. One session was singled out, as it had a noticeable impact on her partner's behaviour towards their daughter, "I think he's done really, really well, to be honest. He's surprised me and I think he's surprised himself. He's opened up a lot more to me. He's been in tears when he's come home, I mean I've been with him eight years, I've never seen him cry or anything. After one session at the course he came home, he apologised to my older daughter. I think it was the child thing that they had to do and pretend like they were in darkness and asleep and people shouting, just put themselves in the child's point of view, if you know what I mean? I think that really done him in and made him realise and he come home and he was in tears and he apologised, like I say, to my older daughter and he's apologised to me a million times. Yes, I think it's a great eyeopener, I think it's a wonderful course."

(Interview with survivor, 004)
Another client identified the children's module of the programme as the point in which he noted changes in himself. He reflected on the transgenerational effect of his behaviour, linking to his own adverse childhood experiences, and understanding of how it may impact his own children,

"Going back to the question you said about, '
At what point on the programme did you feel a change?' For me it was when we covered the children's module, because you learn about your own childhood. You have a real look at yourself, and your own upbringing, and the negative behaviours that you took from that, albeit learned. The facilitators didn't allow you to blame yourself, or your parenting, but to look at the negative aspects of your own childhood experiences, and how that then progressed to your behaviour towards women, towards your partners, and towards your children. Ultimately I could see that I did not want history to repeat itself, however it already had started."

(Client 4, Focus Group Discussion 1)
In their discussion around change, many focussed their attention on spending time and repairing relationships with children. However one respondent goes on to acknowledge that, even though children may be the motivation, spending time with them may not necessarily be the best thing for them, One survivor also points out that her former partner had focussed more on improving his relationship with his children than their relationship,

Theme 3-Concerns around sustaining new behaviours
Finally, clients iterated the importance of sustaining behavioural change, as conveyed through the final theme. To this end, one client considered the potential role of mentors to help sustain change, and to support others to reinforce ideas and behaviours, stating that he was willing to act as a mentor for others who had taken part in the group programme, For this client, even though he felt his behaviour could revert back to the way it was, he shows self-awareness; even if new behaviours are not fully embedded, awareness around behaviour has somewhat improved. This concern around relapsing behaviours, once the support of the group was removed, was echoed by another client,

Discussion
This study provides unique evidence on the outcomes of behaviour, reasons for engagement, as well as the types of abuse committed by clients on a community perpetrator programme by using a triangulation design of various data sources including police data. At the individuallevel, the data provided through the IMPACT Toolkit showed positive changes in both the physical and emotional behaviours related to domestic abuse after completion of the programme. This is consistent with previous evaluations, which found significant reductions in both physical and sexual violence between baseline and twelve months after programme, taken from the survivors' perspective [17,18]. Within the DAPP, survivors also reported improved relationships, particularly in regards to the clients' relationships with children. The qualitative findings were also supported by police data, showing that just over one in four of the individuals enrolled on the DAPP will go on to commit, or be suspected of, any crime after completion the programme. However, domestic abuse crimes amounted to less than half of these total crimes committed by one in five individuals.
These individual level changes were attributed to the quality and length of the programme. In particular, men felt that the time invested in the programme was necessary to embed new behaviours. This finding is consistent with the evaluation conducted on the Tyneside DVPP, whose programme was specifically designed to be long enough to effect real change in men's abusive behaviour [19]. Furthermore, clients on the RADAR/ADAPT programme cited that both the intensity and peer support were required to facilitate change. In the Mirabel evaluation that examines DVPPs across eleven sites, group work sessions were described as being both informative and useful in promoting change; furthermore the one-size fits all was considered the most acceptable approach [17]. Consistent with findings here, input from both facilitators and other men within the group context enabled this behavioural change process.
As found in previous evaluations, the children's module was often mentioned by clients, in particular the sessions where clients were asked to put themselves in the position of children living with domestic abuse [17,19]. The power-control wheel was also frequently mentioned by clients as a useful framework for conceptualising abusive behaviour. The most cited motivating factor for men to join a programme was a willingness to improve relationships, particularly with their children. However, motivation is inherently individual, and there were a number of clients that engaged with a programme who did not have children. As is personal development, with different men being at different stages in their journey towards addressing their own behaviour [20]. According to the transtheoretical model, individuals could be at any stage within the behavioural change process. [21] Some participants had taken steps to initiate change before even starting the programme; for example, by meeting counsellors or therapists. As with other models of behaviour change, change is not a linear process, and some individuals may regress to old behaviours. The RADAR/ADAPT programme is specifically designed to reinforce notions around positive behaviour, moving them towards positive action. However, our results showed that one in five will go on to be a suspect or convicted of a domestic abuse crime after completion. This suggests that, for some individuals, it is difficult to embed new behaviours. Evidence shows that lasting behavioural change can take between six months to five years, supporting the need to maintain new behaviours beyond the end of the programme [22].

Limitations
When interpreting these findings, it is important to acknowledge the following limitations. Firstly, the outcome data was limited to small numbers (n = 34). The IMPACT Toolkit was relatively new to the DAPP, and therefore it took some time to implement its use, which led to missing data on outcomes. Secondly, the end of intervention questionnaire was only completed by clients of the group, and not by survivors (partners or former partners), due to the practicalities around doing this. It was therefore not possible to triangulate these results with the survivors' perspective. Some questions in the before/ after questionnaire assessed the insight of the perpetrator, and how his behaviour affected his partner. It would therefore be more accurate to evaluate impact of harm through the survivor's responses. Thirdly, the RADAR/ADAPT programme was specific to the domestic abuse experienced between intimate partners within heterosexual relationships. The implications are that a number of individuals may not find the programmes appropriate to their context, particular those within lesbian gay bisexual transgender queer (LGBTQ) relationships. Finally, the police data included information on those who had completed the programme only. This meant that a comparative analysis, of those who engaged versus those who did not, could not be carried out which would of otherwise provided valuable information.

Implications
Given that children were a strong motivation for completing a programme, it seemed almost paradoxical that there were no specialist services made available for children within the DAPP model. Consistent with other evaluations on DVPPs, there is a need for a more dedicated support service for children of men on DVPPs, as this is an opportunity to safeguard children. Recent national strategies advocate for early prevention and intervention for domestic abuse, addressing issues earlier in the cycle [17-19, 23, 24].
Although there is compelling evidence to suggest that men changed their behaviours following the programme, police reoffending data suggests that, for a minority of individuals, more work is required to fully embed positive behaviours. A mentoring service may support such aims.
To further support evidence on behavioural change, long term outcomes related to survivor harm should be measured for example, through a short questionnaire filled out by current or former partners. As suggested previously, follow-up periods to DVPPs may be too short. [25] Both nationally, and internationally, there is a need to demonstrate the long term outcomes related to programmes, and to triangulate data with survivors' views. Given that survivor engagement may be poor, shorter versions of questionnaires undertaken by a phone interview may offer a more practical solution [20].