The current cluster randomized trial examines the efficacy of embedding two different parenting interventions within child protection services for young children (0 to 4) who have been exposed to domestic violence and who are at moderate to high risk for recurrent experiences of maltreatment. Interventions are "embedded" in recognition of the significant role played by child protection case workers in identifying families in need of intervention, referring/engaging families in intervention, and being able to use information resulting from intervention (e.g., reports from the intervention program, observations of parenting behaviour changes made as a result of intervention) to improve decision-making in their child protection practice. Thus in this trial, ongoing child protection case workers for families are randomly assigned to receive professional development training, supervision support, and priority client access to parenting interventions in the following four conditions: a) embedded mother-child dyadic intervention (Mothers in Mind); b) embedded fathering intervention (Caring Dads); c) both mother-child dyadic and fathering intervention; d) service as usual.

Mothers in Mind (MIM) is a dyadic mother-child intervention aimed at preventing child impairment resulting from exposure to domestic violence. Intervention focuses on increasing mothers' awareness of the impact that exposure to family violence/trauma may have had on their infants and themselves as mothers, helping identify and promote positive parenting skills such as sensitivity and responsiveness to infant needs, promoting parental competence and emotional closeness and decreasing mothers' social isolation. Mothers in Mind uses an attachment and trauma-informed psycho-educational process approach in 12 weekly sessions (10 group and 2 individual).

Caring Dads (CD) aims to prevent recurrence of child exposure to domestic violence by intervening with fathers. Caring Dads includes 15 group sessions, an individual intake, and two individual sessions to set and monitor specific behaviour change goals. Major aspects of innovation in the Caring Dads program include the use of a motivational approach to engage and retain men in intervention, consistent emphasis on the need to end violence against children's mothers alongside of improving fathering; program content addressing accountability for past abuse; focus on promoting child-centered fathering over developing child management skills; and a model of collaborative practice with child protection.

Hypotheses are posed for differential outcomes among child protection workers (level of randomization) and for children who are the subject of the child protection referral (nested within workers). At the level of the individual child (primary outcome) it is hypothesized that there will be lower rates of re-referral for children of families on the caseloads of child protection workers assigned to the embedded CD, MIM and combined intervention than for those on the caseloads of workers in the service as usual condition. At the worker level (secondary outcomes), outcomes are hypothesized in two areas: 1) worker skill in conceptualizing risk and need in cases of child exposure to domestic violence and 2) increased self-efficacy for referring to and collaborating with embedded interventions. Specifically, we hypothesized that following training and at 12-month follow-up, workers in the CD/MIM intervention and combined CD and MIM condition will have greater case conceptualization skills in responding to hypothetical cases as compared to workers in the treatment as usual condition. We further hypothesize that assignment to an intervention condition will lead workers to report greater self-efficacy for collaborating with embedded parenting interventions than workers in the treatment as usual condition post-training and at 12-months follow-up.

Study focus is the promotion of healthy outcomes in young children (0 to 4) whose exposure to domestic violence (DV) has been substantiated by child protective services. Young children are particularly vulnerable to damage resulting from exposure to DV. Research in developmental neuroscience has shown that infancy and toddlerhood is a time of greater plasticity of the brain and of sensitive periods for the development of a number of core cognitive, emotional, social, and self-regulatory capacities. The public institution with the greatest opportunity to promote better outcomes among very young children at high risk for compromised mental and physical health development as a result of exposure to domestic violence is child protective services. Exposure to domestic violence is one of the most frequently substantiated forms of child maltreatment experienced by Canadian children. As with other forms of maltreatment, very young children experience disproportionately high levels of victimization. Moreover, because infants and toddlers are more often in the presence of their mothers than older children, their exposure experiences are more likely to be direct (i.e., witnessing violence) as opposed to indirect (i.e., hearing or knowing about violence).

Unfortunately, there is ongoing concern about child protection response to domestic violence and about the availability of interventions to address this issue. Interventions are needed in two areas. The first is interventions that will reduce impairment (i.e., cognitive, social, emotional development) of young child victims of exposure. This issue has been the focus of many academic and policy critiques and the source of considerable tension within and between the child protection and Violence Against Women (VAW) service communities. One tension is that mothers, on whom children rely for sensitive contingent responding to traumatic events and traumatic reminders, are also victims of DV. There is a robust relationship between DV victimization and symptoms of trauma and depression, and between maternal symptoms of depression and trauma and elevated levels of maternal intrusiveness, hostility, and non-responsiveness to young children. Domestic violence victimization is also a strong risk factor for mother-perpetrated maltreatment and many of the more complex child protection cases present with a combination of risk due to ongoing concerns about fathers' perpetration of domestic violence and about mothers' DV- or mental health-related neglect. There have been numerous calls for better training and greater collaboration between child protection and VAW services at this complex area of practice to avoid retraumatization of women and to promote better outcomes for children in the context of exposure to domestic violence.

Second, a strong child protection response to DV needs interventions to prevent the recurrence of child exposure to violence. Historically, child protection practice in DV has focused almost exclusively on mothers' capacity to take actions to "appropriately protect" their children from violence exposure. There are many problems with this mother-focused strategy of child protection. The more appropriate alternative response is to offer effective services to parents (most often fathers) who have perpetrated DV in their families. Child protective services have been relatively slow to include fathers in their work, though this is changing. There have been numerous calls to continue to change practice in this area in order to make work with fathers a greater part of child protection responses to child exposure to domestic violence.

The current study is designed to test the efficacy of providing families' child protection ongoing service workers with professional development training, supervision support, and priority client access to embedding two parenting interventions - one for mothers focused on preventing impairment and one for fathers focused on preventing recurrence - into child protection practice. These interventions are "embedded" in recognition of the unique context of child protection practice. Within child protection, families are assigned primary workers (in this case, ongoing service workers) whose job it is to work with a family to improve child safety to the point that the child is safe enough to close the file. Surprisingly, referral to parenting intervention is seldom part of child protection practice - most often, workers rely on their own work with families to prompt change. However, when parenting interventions are suggested, this referral is often linked to child protection plans; i.e., families are not really choosing to access these interventions on their own accord. [There is nuance to this, as families are not legally mandated to attend. Yet given the power difference between the child protection worker and the family, a strong suggestion from a worker that a family access an intervention program carries substantial weight. Families retain the ultimate choice about whether or not to participate, but this choice cannot be understood as fully independent and voluntary]. Moreover, parents' success (or failure) in attending these programs and in making changes to their parenting can have implications for the length, intensity, and nature of ongoing involvement of the family with child protective services. In this context, it is not feasible to have workers "suggest" that a family access an experimental intervention program and then be randomly assigned to receive, or not receive, this program. Moreover, given this context, improving outcomes by providing parenting intervention requires that change occur first at the child protection worker level. Specifically, workers need to be able to identify families in need of intervention, be successful at referring/engaging families in intervention, and be able to use information resulting from intervention (e.g., reports from the intervention program, observations of parenting behaviour changes made as a result of intervention) to improve decision making in their child protection practice. Improved child outcomes also hinge on having interventions that are themselves efficacious. Because of the embedded, systems context of parenting intervention for child protection client, and because the first point of change is worker referral, the appropriate level of experimental manipulation is the child protection worker, not the individual families. Accordingly, this study makes use of a cluster-randomized design, where child protection workers are assigned to additional training, consultation, and support in making referrals to mothering, fathering, or both interventions as compared to practice as usual. Also relevant to considering design is the fact that child protection workers who will be the subjects of this CRT are nested in teams of 5 or 6 under a supervisor. Supervisors are responsible (generally speaking) for providing clinical supervision, administrative management, and leadership to their team of workers. As such, supervisors should be understood as potential "gatekeepers" (i.e., someone who may be called up to protect group-based interests that are affects by enrollment in a CRT). Given this organizational structure, the most practical and realistic design is to nest randomization by teams. Thus, workers in any one team will all be assigned to the same condition.

Parenting Interventions, Child Protection and Clinical Equipoise In conducting a clinical trial, it is also necessary that genuine uncertainty exists in the relevant expert community about what therapy(ies) are most effective. Here, the relevant question is whether referral to a formal mothering/fathering intervention program is more efficacious than providing regular ongoing child protection service, which consists of in-home visits and individualized problem-solving with families. There is genuine uncertainty about this question in the area of both interventions for mothers and intervention for fathers.

In terms of intervention for mothers, a major question is whether a "therapeutic" parenting program is efficacious when offered as part of child protection service where the typical provisions around client-therapist confidentiality are replaced by an agreement for collaborative information sharing between the intervention program and the child protection worker. This issue has been written about frequently and is one of the sources of tension between the CAS and VAW sectors. Also of concern is whether this difficult to engage group of mothers can be engaged in services when they are embedded in child protection practice.

The specific program to be offered in this trial is Mothers in Mind. The Mothers in Mind program was developed in response to a recognized service gap for mothers who had experienced abuse/trauma, were showing abuse-related deficits in parenting, and had children under the age of four. This program is built on the foundation of research showing that dyadic parent-child responsivity based approaches are efficacious for improving outcomes for children in at-risk circumstances, including infants born prematurely, parented by depressed mothers, dyads who have been trauma exposed, and who are in anxiously attached. There is preliminary evidence in support of the value of MIM. A 2013 intervention study found that both facilitators and management identify that they are satisfied with the support provided by CDI in the areas of consultation and training. Pre- to post-program self-report evaluation data suggest that the MIM program is specifically helpful to mothers initially presenting with challenges in parenting. Specifically, in the subgroup of mothers presenting with high levels of challenge, completion of the MIM program was associated with significant positive changes in mothers' isolation, attachment, and sense of parenting competency (no other dimensions of outcome were assessed).

In terms of fathering intervention, there is a great deal of concern in the literature about the efficacy of any group intervention aimed at reducing abuse recurrence, especially domestic violence. A recent meta-analysis concluded that the research evidence on programs aiming to change men's perpetration of domestic violence is insufficient to draw conclusions about effectiveness. Specifically, Smedslund et al. noted in 2011: "This does not mean that there is evidence for no effect. We simply do not know whether the interventions help, whether they have no effect, or whether they are harmful" (p. 8). Given the state of evidence, experts would certainly be divided on their opinion about whether a trial of this nature will show any advantage over service as usual.

The specific program to be examined in this trial is Caring Dads. Caring Dads aim to reduce recurrence of domestic violence by working with fathers to increase their healthy, supportive, and non-abusive engagement with their partners and children.

Caring Dads is currently considered a "promising practice" for addressing child maltreatment. Preliminary research on Caring Dads, using a comprehensive evaluation framework, established that Caring Dads addresses a need in communities, can be implemented in a way that is acceptable to clients and stakeholders, and matched, in its underlying theory, the characteristics and needs of most referred. Subsequent examination of Caring Dads using a pre- to post-research design showed that intervention is associated with changes in fathers' over-reactivity to children's misbehaviour and respect for their partner's commitment and judgment, with results being statistically significant, medium in size, moving mean scores into the normative range. Interim findings from independent research on Caring Dads being carried out in the UK are also promising. Based on data from 204 fathers, 72 partners, and 22 children, McConnell reports that completion of Caring Dads is associated with pre- to post-group reductions in parenting stress and in level of hostility, indifference, and rejection as reported by fathers and reductions in domestic violence victimization (emotional abuse, isolation, violence, injury, use of children), depression, and anxiety as reported by mothers. Changes in identified domains persist over six-months and are well in excess of changes made by comparison group fathers over a similar time period. Finally, a small ongoing study by the Child Welfare Institute and K. Scott in Toronto finds that, consistent with Caring Dads' model of collaboration between group co-facilitators and child protection workers, enrolment in Caring Dads is associated with substantially higher levels of contact between men and their families' child protection workers and lower rates of re-referral for men enrolled in the program as compared to those referred but on a wait-list for service.

Both the MIM and CD interventions will be embedded as a collaborative intervention across child protection and community service. Child protection workers will be directly trained by staff of the MIM and Caring Dads program on the nature of the programming provided. Cases potentially eligible for this program on the basis of referral and ongoing concerns will be flagged. Moreover, child protection staff will have ongoing consultation around the complexity of how to interpret and effectively respond to women's joint status as survivors of DV and mothers ensuring the safety and well-being of their children and on how to best intervene with fathers. Other aspects of embeddedness include ongoing communication through intervention and collaborative planning (child protection and community agency) and agency-CAST co-facilitation of intervention groups.

Duty of Care Considerations. The child protection workers and supervisors being recruited into this study have a duty of care to the clients on their caseload. As part of their job, they need to make decision as to the best services and interventions to provide to their clients. Workers and supervisors will be fully aware that they are participating in a study where we are examining the effect of embedded parenting interventions. They will also be made aware that there is genuine uncertainty about the value of these interventions and will be cautioned about the possibility of therapeutic misconception. Depending on the trial arm, workers will be trained to better assess and recognize issues in the parenting of mothers and fathers of young children and they will be given information about intervention programs that have promising evidence of success. In addition, cases potentially eligible for referral to these programs will also be "flagged" for workers. However, it is important to note that there will be no (explicit or implicit) mandate that flagged clients will be referred to these intervention programs. Workers and supervisors will need to continue to use their judgment about the ultimate appropriateness of referral for individual clients. In addition, workers not assigned to a particular intervention arm (e.g., those assigned to service as usual) will not be prevented from making referrals to CD or MIM. Although workers will again be aware of the uncertainty of outcomes of these interventions (and the possibility of therapeutic misconception) and of their assignment in the trial, we will respond to their professional judgment about referral. If space is available for clients of these workers, they will be offered intervention regardless of the fact that this will weaken the intervention design. However, we anticipate that this will occur in a low number of cases because survey of practice reveals that few workers make referrals to parenting interventions and because within the context of this trial, we do not expect to have many "empty" spaces in intervention.

Conditions and Hypotheses

As explained earlier, because the focus of this study is on changing the ways in which child protection workers embed intervention into ongoing child protection family service, the appropriate level of experimental manipulation is the child protection worker, not the individual families. Outcomes will be examined at two levels - the level of the child protection worker and the level of the young child identified as being in need of the service of the Children's Aid Society (CAST). This trial will have four condition arms: a) embedded mother-child dyadic intervention (Mothers in Mind); b) embedded fathering intervention (Caring Dads); c) both mother-child dyadic and fathering intervention; d) service as usual. Hypotheses are outlined in detail as follows:

Worker level Case Conceptualization H1: Professional training on embedded interventions (i.e., training, ongoing consultation and supported referral) will increase the workers' capacity to conceptualize risk and needs in hypothetical cases typical of those open to child protection services as a result of child exposure to domestic violence. Specifically, workers will demonstrate greater capacity to identify intervention needs relevant to fathers'/mothers' attitudes and behaviors; will be more likely to recommend interventions relevant to addressing fathers'/mothers' attitudes and behaviors; will be more likely to identify specific, concrete behavior changes in fathers/mothers as part of the rationale for case closure; and will increase the extent to which workers take intervention progress, or lack of progress, into account when making decisions about case closure in responding to hypothetical case descriptions. Effects will be evident following training and at a 12-month follow-up point in case conceptualization around mothers (for the MIM and combined intervention groups) and fathers (for the CD and combined intervention groups) as compared to workers in the treatment as usual condition.

H2: Workers randomly assigned to an embedded intervention condition will have greater self-efficacy for referring to and collaborating with intervention programs. Effects will be evident following training and at a 12-month follow-up point. Specifically, workers in the CD and combined condition (i.e., conditions 3 and 4) will report greater self-efficacy for referring fathers and collaborating with fathering interventions than workers in the treatment as usual condition and workers in the MIM intervention and combined condition (i.e., conditions 2 and 4) will report greater self-efficacy for referring mothers and collaborating with mothering interventions than workers in the treatment as usual condition.

Child Level Rates of re-referral for child protection concerns H3: There will be lower rates of re-referral for children of families on the caseloads of child protection workers assigned to the embedded CD, MIM and combined intervention than for those on the caseloads of workers in the service as usual condition.

H4: Young children on the caseloads of child protection workers assigned to the embedded CD, MIM, and combined intervention will have fewer emotional and behavioural symptoms at one year follow-up than children of workers in the service as usual condition.

Inclusion Criteria:

• Did the intake process substantiate child exposure to domestic violence as a primary or secondary code (Yes)
• Was the child's father (biological, step, or common-law) identified as a perpetrator of this abuse(need not be exclusive perpetrator)? (Yes)
• Is at least one of the investigated children four years old or younger (yes for MIM comparison)?
• Does the maltreated child reside with his or her mother or father or with his or her mother with ongoing regular contact with his/her father? (Yes)
• Are there continuing and insufficiently addressed concerns about the potential recurrence of domestic violence by father (Yes)
• Are there continuing and insufficiently addressed concerns about problems in the mother-child relationship? (Yes)

Exclusion Criteria:

• The child's mother and father are unable to communicate in English
• Given the opportunity, this father could not participate in a group-based intervention (reasons are documented. Examples would be severe mental health/substance use issues, severe problems with cognitive functioning, a work schedule that requires him to be regularly out of the region)
• Given the opportunity, this mother could not participate in a group-based intervention? (reasons are documented. Examples would be severe mental health/substance use issues, severe problems with cognitive functioning, inability to arrange with work schedule)