September 10 2002 www.restorativepractices.org

Restorative Practices E FORUM

Long Term and Immediate Outcomes of Group Conferencing in Washington State (June 2001)

NANCY SHORE, JUDITH WIRTH, KATHARINE CAHN, BRIANA YANCEY, KARIN GUNDERSON

ABSTRACT in the eighties and nineties, public opinion situation is more complicated than the This article presents the findings of a retrospec- began to blame this focus on family preser- smooth easy swing of a pendulum between tive study of 70 family group conferences (FGC) vation, and called for renewed attention to family preservation or . conducted in Washington State. These 70 FGCs safety. The challenge for the child welfare system is addressed the well-being of 138 children. The The passage of the Adoption and Safe Fami- to move practice, policy, and thinking off within the evaluation were primarily re- lies Act of 1997 (ASFA) was seen as a push to the restrictive paradigm of the pendulum ferred by units rather than investiga- tive units and involved cases that had been in the the family preservation end of the arc, pri- swing. Social workers and families need to child welfare system for over 90 days. Families oritizing the safety of the children. Policies work together to create a plan that provides were invited to participate in the decision-mak- and programs shifted from family preserva- ing process, engaging both the maternal and pa- tion to pushing for timely safe permanent ternal sides of the family with greater success than placements for children. Adoption was iden- Social workers and standard case planning approaches. Children who tified as the most-rewarded strategy to re- families need to work had a conference experienced high rates of re- duce the high rate of children drifting in the together to create a plan unification or kinship placement, and low rates foster care system, and incentives were of- of re-referral to CPS. These findings generally fered to states for each adopted child. As that provides for the remained stable as long as two years post-con- ference. This study, the largest long-term fol- ASFA specified shorter timelines for child's immediate safety low-up study of FGC published to date, suggests to demonstrate their ability to safely care for that FGCs can be an effective planning approach their children, more than one advocate won- and takes into account a for families involved with the public child wel- dered whether parents were becoming no child's long term fare agency, resulting in safe, permanent plans more than "speed bumps" on the way to ter- developmental needs. for children at risk. mination trials. Thinking in terms of a pendulum, however, INTRODUCTION is problematic. It frames child welfare as an for the child's immediate safety and takes into Child welfare policy and practice in the either/or choice of child safety or family account a child's long term developmental United States have been described in terms preservation. Sometimes swift action to ter- needs. In the process, it is also critical to of a pendulum, swinging between child safety minate parental rights is needed due to ex- broaden notions of family to include the and family preservation. The landmark treme circumstances in the birth family. But network of . The social Adoption Assistance and Child Welfare Act for the majority of the children the situa- worker and family must have a fuller range of 1980 (P.L. 96-272) represented a swing tion is more complex. For example parental of motion than the simple back-and-forth towards family preservation where policies prognosis for recovery from addiction hard arc of a pendulum. "Remove" or "reunify" promoted efforts to keep families intact and to assess, and likely will take more time than cannot be the only choices given the com- prevent the placement of children into fos- the child's developmental timeline (as re- plexity of child welfare cases today. ter care (Cole, 1995). As the country flected in ASFA timeframes) requires. In This challenge is particularly critical for struggled with rising foster care placements other situations, the child may have special families of color who are involved with the and a number of high profile child deaths needs that require extraordinary care. The child welfare system. Over one-half of the

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM

African American children (56%), for ex- & Nixon, 1999), Australia (Swain, 1993), More work is needed with larger sample sizes ample, receiving child welfare services are in Canada (Immarigeon, 1996) and in parts of and in other settings to gain a greater un- foster care placements - twice the percent- the United States, including Colorado, New derstanding of long-term outcomes for age for white children (Children's Defense Hampshire, New York, Massachusetts, Cali- families. The present evaluation addresses Fund, 1999). Despite no higher levels of fornia, Oregon, and Washington. An un- this gap in our understanding of the effec- abuse African American children are more derlying philosophy of the FGC model is that tiveness of the FGC model by reporting both likely to be removed (Morton, 1999; US extended families have the commitment, re- the immediate and long-term outcomes for Department of Health and Human Services, sources and capacity to create safe and car- an ethnically diverse group of 138 children 1999), experience long stays in out of home ing plans for their children. in Washington State. care, receive fewer services (Barth, 1998; Descriptive studies, primarily focusing Close, 1983; Courtney, 1996; Olsen, 1982) upon process measures and immediate re- An underlying philosophy and wind up as legal orphans with no com- sults, show that FGCs engage more family mitted permanent home (Kemp, 1999). For members than other case-planning meth- of the FGC model is that these and all children, a broader definition ods, result in high degrees of family and extended families have the of family and a commitment to engage fam- professional satisfaction, and expand the commitment, resources ily in the case planning process is one strat- quality of support available to families who egy that could help alleviate these dispropor- have participated (for a review, see Lupton, and capacity to create safe tional numbers and increase the likelihood 1999). Findings from child welfare stud- and caring plans for their that services provided would be more cul- ies where there was not a FGC provide sup- turally appropriate. port for the importance of active family children. The practice of family group conferencing involvement. Gleeson et al (1997), for ex- (FGC) is one way to get off the pendulum ample, found that an absence of active FAMILY GROUP CONFERENCING IN THE STATE and address the safety of the children in the family involvement in case planning and OF WASHINGTON context of a permanent connection with decision-making can create a barrier to families. As practiced in Washington State, achieving permanence. FGC has been practiced in Washington State it can provide for both immediate safety, and since 1997 when two Division of Children Prior to claiming the success of FGC as an and Family Services (DCFS) managers con- intervention, however, there is a need to tracted with the University of Washington to "Remove" or "reunify" understand whether these immediate re- implement a statewide pilot demonstration cannot be the only choices sults are sustained over time. We need to of FGC. The original FGC pilot program is know, for example, whether the place- given the complexity of now standard programming with long-term ments that appear stable six months after investments in five of the six administrative child welfare cases today. the conference continue to be stable and regions, each of which have allocated regional safe environments for the child in the budgets to fund one to six FGC facilitators long-term. Few FGC studies have exam- per region. The state child welfare agency long term family permanence and parental ined longer-term outcomes. One excep- connections. A description of the model, contracts with the University of Washington tion to this is the work of Pennell and to convene facilitators from around the state including its origins, will be presented fol- Burford (2000) with families experienc- lowed by an evaluation that looks at both the on a monthly basis in order to support the ing in Newfoundland exchange of best practice knowledge and to immediate and long-term benefits of FGC and Labrador. In follow-up interviews and for 70 families. provide direction for the future of FGCs. progress reports that took place an aver- At a later date, the Stuart Foundation of WHAT IS FAMILY GROUP CONFERENCING? age of one year post-conference, the au- Washington and California added funds for thors found that FGCs did a better job an evaluation component. Family group conferencing (FGC) is a par- than regular case planning approaches in ticipatory approach to case planning that was promoting family unity, increasing safety In the state of Washington, FGC consists originally developed by the Maori people of for all family members, and reducing re- of a three-stage process where-by families New Zealand, in response to concerns that ports of child maltreatment and mother/ assume a central role in planning for their the child welfare system was removing Maori wife abuse. The study found that overall children within the mandated authority of children from their homes and cultural ties levels of abuse had decreased significantly the child protection agency. The FGC at a disproportional rate. Based upon the for the families involved with the project process used in Washington State, as de- success of this approach in New Zealand, and increased moderately for the compari- scribed in the next section, reflects those FGC has been utilized as a case planning son group. used in other parts of the country and in- approach in the (Lupton ternationally.

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 2 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM

STAGES IN A FAMILY GROUP CONFERENCE immediate needs, such as treatment plans or sure service delivery was cited as a major con- 1. Referral and Contacting Family visits with family members. cern (Mason Report as cited in (Lupton & Nixon, 1999)). A case carrying social worker refers a family c. Reconvene with Social Worker to an FGC facilitator. The facilitator then After the family has developed their plan, the For more detailed descriptions of the pro- works with the child's parents or relative FGC facilitator and the referring social cess or history of FGC see (Hudson, Mor- caregiver to identify other extended family worker reconvene with the family. The fam- ris, Maxwell, & Galaway, 1996,Hassall, 1996 members and support people who should be ily presents their plan to the social worker, #196) (Robertson, 1996b); and (Lupton & invited to the conference. During this pe- who is responsible for approving the plan. Nixon, 1999). riod the FGC facilitator also contacts ser- METHODOLOGY vice providers who can provide pertinent information regarding the well-being of the Study Sample and Procedures child. At this stage, the facilitator's task is to After hearing all the The findings reported here were generated prepare both the family and service provid- information, the family through a participatory evaluation conducted ers for their role in the conference. For ex- in close partnership between DCFS and the ample, the facilitator explains to the family and their support network University of Washington. During the pro- members that their charge is to develop their meet privately to develop a cess of the evaluation, members of the Wash- own family plan that addresses the safety and ington State FGC facilitator team provided well-being of the child. Service providers are plan aimed at assuring the critical guidance and insight including sug- told to provide information to family mem- well-being of the child. gestions for design features, identification bers, but not to express an agenda or an out- of questions for analysis, and assistance with come. The only exception to this is when the interpretation of project findings. This there is a serious safety concern. For ex- The primary criterion for approval is the collaborative process strengthened the valid- ample, if a plan to return the child to a par- safety and well-being of the child. Previous ity of the evaluation. ent would not be approved due to safety con- studies have found that the majority (in most The sample for this evaluation included all cerns, the referring social worker is asked to studies as high as 90-95%) of families are conferences that were conducted by two FGC be explicit about any "bottom line" safety able to identify a plan, and that these plans facilitators in a suburban and rural area limits. are approved by the referring social worker north of Seattle, Washington. These two fa- 2. The Conference (Crowe & Marsh, 1997; Simmonds, Bull, & cilitators tracked descriptive data and out- Martyn, 1998);(Lupton & Sheppard, 1999). a. Information-gathering come information on their FGCs dating back If the plan is approved, it becomes part of to the start of the project in May of 1997. The facilitator convenes the meeting at a lo- the child's record and is usually sent to the Because the intent of the study was to explore cation selected by the family and the meet- Dependency Court judge. long-term outcomes, only cases at least 6 ing begins with introductions and a state- 3. Follow-up months post-FGC (occurring prior to No- ment of purpose by the facilitator. The pro- Follow-up is an important phase of the FGC vember 1999) were included in the analysis. viders then proceed with sharing informa- The evaluation consequently reflects out- tion with the family. This can include evi- process as the plans must be monitored to ensure that children remain safe, families are come information on 70 families with a to- dence leading to the concerns for the child tal of 138 children. and family well-being, a description of the receiving adequate support, and that services services currently provided and other com- are being delivered. For some families, fol- There were two sources of information for munity resources, supports the child welfare low-up may be in the form of a second FGC. the evaluation. The first source of informa- agency can provide, explanations of the per- Conceptualizing the FGC process as three tion, the family plans, were prepared by the manency planning options, and if appropri- phases emphasizes how specific efforts are families at the conference, outlining the ate, an overview of how the dependency court required to 1) prepare for the conference, steps families agreed upon to assure the well- system works. 2) facilitate the conference, and 3) provide being of the child. A content analysis of the support to the families after the conference. family plans provided information on the b. Private Time This conceptualization of the FGC process immediate outcomes of the conferences. After hearing all the information, the fam- moves us beyond viewing the conference as a Attached to these plans were face sheets that ily and their support network meet privately single act that can be successfully performed FGC facilitators routinely completed after to develop a plan aimed at assuring the well- without careful preparation work or adequate the FGC. These face sheets included demo- being of the child. Often the focus is to de- support after the family meets. In looking at graphic and departmental information; for velop a safe permanent plan, but other times a critique of the New Zealand experience with example information on what unit referred the family may choose to focus on other more FGC, lack of follow-up to monitor and en- the family for an FGC. The second source

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 3 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM was a database of outcomes kept by the FGC placement options after the FGC, and ethnicity. Our analysis was not able to ac- facilitators. The database included case-by- whether the plans reflect variations in fam- count for the potential ethnic variations case information, such as whether the child ily and cultural approaches to care-giving within families. returned to care after the conference. The and problem-solving. Based upon practice Family Member Participation Was High information from the database allowed us to experience, it is believed that families will address the most central questions in this tap into their own resources, with the sup- Participation at the FGC was considered paper, whether the immediate outcomes re- port of the agency, to create a permanent an indicator of whether extended family sulted in sustained long-term benefits for plan that accounts for the well-being of the members mobilized around the planning the children. child. There are three primary long-term for the child's well-being. Data show that each conference drew a high number of The University evaluation staff gained access questions that will be addressed within this study: (1) were there any substantiated CPS family members, and those family mem- to the database and the family plans only af- bers out-numbered service providers. ter identifying client information was re- re-referrals, (2) were the placements stable, and (3) did the case close? Across the 70 FGCs, there were 589 fam- moved, thereby assuring protection of cli- ily members and 361 service providers in ent confidentiality. The FGC facilitators es- RESULTS attendance, resulting in an average of 8 tablished a coding scheme that allowed for Who participated in the FGCs? family members and 5 providers at each the database and plans to be linked while conference. Family members include fic- maintaining the clients' anonymity to the Families Were Primarily Referred Through tive kin, friends of the family, or anyone non-DCFS evaluators. Permanency Planning Units the family identified as a support person, The extent to which we could answer the Of the 66 families where information was such as a pastor. Types of providers attend- evaluation questions was limited at times by available, 74% were referred from CWS, ing the conferences include school coun- the availability of information. The content 23% from Child Protective Services (CPS), selors, therapists, and foster parents. of the database and the face sheets were col- and the remaining 3% from Tribal Indian Family involvement was further explored lected prior to the conception of this evalu- Child Welfare workers who provide both CPS to learn about the degree of participation ation, thus limiting what questions could be and CWS services. According to the FGC by the maternal and paternal sides of the addressed within this evaluation. In some facilitators, many of the families were con- family, as well as for the presence of chil- instances information was collected using the sidered "challenging" cases where a plan for dren at the conferences. For 57 of the cases child as the unit of analysis, while for other the children had not yet been identified. Of within this study, information was avail- variables the family served as the unit of note, children who are involved with CWS able regarding maternal and paternal par- analysis. Furthermore, there was a problem typically are receiving permanency planning ticipation. For some of these "missing with missing information. As the number of services and have been in out-of-home care cases," only the plans identified at the fol- conferences increased, the facilitators' pri- more than 90 days. Substance abuse and low-up FGC were available for analysis. ority understandably focused on preparing neglect were the two primary concerns that For purposes of consistency, it was decided and convening conferences rather than brought these families into the child welfare not to include these plans in the study. tracking all the conference data. Through- system. Across these 57 cases, there were 263 ma- out the report care will be taken to indicate The Diversity of Families Was Similar to That ternal and 163 paternal relatives, result- the number of cases where information was in the Agency as a Whole ing in an average of 5 maternal and 3 pa- available and whether the unit of analysis was The diversity of families within our sample ternal relatives attending a conference. the child or the family. reflects that of the general child welfare FGC's commitment to family involvement EVALUATION QUESTIONS population in this region, with two excep- in case planning was reflected in the tions. Within the sample, 59% of the fami- model's ability to creatively accommodate Through the descriptive data the follow- different family situations. For example, ing questions are addressed: (1) which lies were Caucasian, 23% were Native American, 11% African-American, 4% at several of the FGCs, parental participa- child welfare units referred families, (2) tion occurred via the use of speakerphone what was the cultural diversity of the fami- Hispanic, and 3% Asian. Proportionally, there are fewer Caucasian families and with an incarcerated . Sixteen chil- lies, and (3) who participated at the con- dren were listed as being physically present ference, both in terms of the composition more Native American families in the sample compared to this region's represen- at the conference. Other children's voices of the family support network and the were shared, however, through their let- presence of service providers. The imme- tation of children in out-of-home care (Washington Permanency Summit, 1999). ters being read at the conference or by the diate outcome questions include whether participation of their therapist or coun- the families' plans were approved by the Of note, information was only available regarding the family's ethnicity as a whole, selor. Information was not available re- social worker, whether there was an in- garding the age of the children. crease in reunification or relative care while the regional statistics describe child

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 4 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM

What were the immediate outcomes? TABLE I. WHERE DID THE CHILDREN GO? (N=114) The Conference Resulted in Family Plans in Pre-FGC placement Post-FGC placement Most Cases Living with parent(s) 20% 43% A family plan, outlined during the family's Living with relatives 55% 31% private time and approved by the caseworker, Living with non-relatives 25% 9% is the primary goal of the FGC. For 97% of Tribal jurisdiction 0% 4% the children (n=134) a family plan was iden- Plan not achieved N/A 13% tified. For three of the other children, the family chose to reconvene in order to de- these plans, families listed a mental health table below lists where the children were placed cide upon a plan. For only one child was the service such as a psychological evaluation or pre- and post-FGC. Children placed in fos- family unable to arrive at a plan. All of the counseling. The second most frequent ser- ter care, detention, and treatment facilities identified plans were approved by the social vice listed was substance abuse treatment or for special needs were considered to be living worker indicating that the plans met agency prevention (61%), which included AA meet- with non-relatives. The greatest shifts can be standards for child safety and well being. ings or substance abuse evaluations. The seen with the number of children living with These findings correspond to the results of third most frequent service was behavioral parents and relatives. After the conference other FGC studies (Maxwell and Robertson, interventions (61%) that included services there was a surge in the percentage of chil- 1991; Paterson and Harvey, 1991; Renouf, such as anger management, domestic vio- dren living with their parents and a decrease Robb, and Wells, 1990; all as cited in lence services, parenting, and stress manage- in the number of children living with rela- (Robertson, 1996a). ment classes. The last main grouping per- tives. This shift suggests that children were Family Plans Combined Both "Traditional" tained to housing resources (30%). Other reunified after living with a relative. The per- and Family Specific Strategies resources that families identified were edu- centage of children living with a non-relative As the families are empowered to create their cational services, public assistance, Intensive also decreased after the FGC. own plans, the FGC process is more likely to Family Preservation Services or less-inten- FGCs Resulted in Permanent Plans or Other reflect the variations in family and cultural sive family preservation services, childcare, Positive Immediate Outcomes public health, and health. approaches to care-giving and problem- Public Law 96-272, the Adoption Assistance solving. The family plans created at the FGC In addition to the more "traditional" ser- and Child Welfare Act of 1980, defines per- were reviewed to determine whether they in- vices, families also identified resources that manence as "the least restrictive (most fam- cluded strategies that were both representa- tapped into their own strengths. In all of the ily-like) setting available and in close prox- tive of the types of traditional services seen 57 plans reviewed, at least one family-driven imity to the parents home, consistent with in case plans where a FGC did not occur, support was listed. These services included the best interests and special needs of the and that were also unique to the family. Only family members providing transportation, child." This permanence can be codified by 57 plans, out of a possible 70, were available financial assistance, supervised visits, con- law as reunification or adoption, or by the for analysis. The more "traditional" services tinuous emotional support, respite care, and agency as guardianship or long term foster long-term placement resources. Other fam- care. Where information was available As the families are ily plans included extended family members (n=129), the FGC resulted in a permanent helping with school tuition, providing fur- empowered to create their placement for 82% of the children. In ad- niture, attending children's extracurricular dition there were two pending and own plans, the FGC activities, and contributing to home im- a relative placement that possibly was in- process is more likely to provements. The plans also identified cul- tended as a permanent living arrangement. tural supports such as sweat lodge healing and For the remaining 15.5% of the children, reflect the variations in church based supports. either the plan was not achieved or it was a family and cultural The Proportion of Children Living With a nonpermanent placement. approaches to care-giving Parent Increased After the FGC It should be noted, however, that not every Information regarding placement location FGC was designed to address issues of per- and problem-solving. prior to the conference was available for 114 manence. For example, at one FGC a treat- of the children. Knowing where the child was ment plan was the issue under discussion. identified by families were categorized into placed prior to the conference is critical in The family specified a timeline for the four main groups: mental health services, order to explore how placement location mother to receive therapeutic services and substance abuse treatment, behavioral inter- changed after the conference. We therefore demonstrate her ability to care for the child. ventions, and housing resources. In 80% of limited our analysis to these 114 children. The The plan was to reconvene at a follow-up

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 5 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM meeting to assess whether it was appropriate The Rate of Re-referral Was Low Over Time the 14 children were placed in relative care. for the child to return home or whether an- An indicator for child safety is whether there Of the 14 children who experienced a change other permanent plan should be pursued. are subsequent founded allegations of abuse in placement, only 3 of these children had a In six other cases, the "non-permanent" and neglect. This study was able to find in- substantiated CPS re-referral after the con- placement was intentional, as the family had formation on 133 out of the 138 children. ference. Although there were no significant identified the placement as part of their sec- Information was not available for 5 of the differences in the rates that children re- ondary plan. Families were encouraged to children due to a transfer to tribal jurisdic- turned to care by the amount of time elapsed identify a secondary plan in case the primary tion, thus taking the children out of the state since the conference date, it is important to plan did not work out. For example, one child welfare management information sys- note that for the 55 children who had a con- youth had a primary plan to return to the tem. Of the 133, nine had a founded CPS ference 2 years ago, only 5 (9%) returned to care of relatives with a secondary plan to re- referral after the conference, resulting in a out-of-home care. Of the remaining chil- turn to group care if the youth's behavior 6.8% re-referral rate. Of the 55 children dren who returned to care, 6 had partici- became too disruptive. who were the focus of the FGC over two years pated in a conference one-to-two years ago prior to the study, there were substantiated and the other 3 six months-to-one year ago. The proportion of CPS referrals on only two of the children. Findings Were Similar in Sexual Abuse Cases Further analysis of the cases that were re-re- Safety is a particular concern for families children living with a ferred show for over half of these children where sexual abuse is an issue because of the parent increased after the the re-referral did not disrupt the family secrecy surrounding the abuse. Because of FGC. plan. The re-referred cases involved inci- these safety concerns, the FGC team was in- dents occurring within five families, as some terested to learn of the outcomes for fami- of these 9 children were . Although lies where sexual abuse was the issue. For Placements Were Designed With Both the an allegation of neglect was substantiated for thirteen of the families (26 of the children) Maternal and Paternal Side of the Family three children, no change in the plan was sexual abuse was identified as one of the risk In this study there was a high rate of pater- deemed necessary after the investigation. For factors calling for state intervention. Family nal presence at the FGCs and a high rate of two other children where there was a re-re- plans were identified for all of these chil- children placed with paternal relatives. Af- ferral, the result was a change to the second- dren at their conferences. After the confer- ter the conference, approximately one-third ary plan that had been identified at the con- ence, 21 of the children either remained in of the children found a home with either ference: group care and relative placement. their parent(s)' home or were placed with a their father or a paternal relative. More spe- The remaining 4 children who had re-re- relative, 4 children went to out-of-home cifically, 24% of the children were placed (or ferrals were placed in foster care, which was care, including one foster parent guardian- remained) with dad, 10% with a paternal not part of the family plan. Information was ship and information was not available re- relative, 28% with mom, 20% with a mater- not available on the frequency of inconclu- garding the final child. There were no sub- nal relative, 6% with both parents, and 12% sive and/or unfounded reports. stantiated CPS re-referrals after any of these with a non-relative. Placement information Placements Were Stable Over Time conferences. that specified the relationship of the As we have reported, the majority of fami- A Slim Majority of Cases Were Closed at the caregiver to the child was available for only End of Two Years 110 of the children. lies were able to identify a placement plan for their children. It is important, however, Closing a case is usually an indication that Did the Children Remain Safe and in Stable to look beyond whether a placement plan was the social worker considers the child to be Placements Over the Long-Term? identified and determine whether or not the in a safe situation, that a permanent plan has The question of long-term safety and stabil- plan remained stable over time. At the time been completed, or that agency involvement ity was operationalized in three ways: 1) were of the study, the majority of children were no longer provides additional benefit or is there any substantiated CPS re-referrals, 2) in the placement identified in the family necessary. Cases were reported as closed for were the placements stable, and 3) did the case plan. Only 14 (10.1% of 137) of the chil- 54% of the 2 plus-year subgroup (n=50), close? These questions were examined for dren experienced difficulties with the in- 34% of the 1-2 year subgroup (n=59), and three groups of cases: those more than two tended primary plan and consequently were 13.6% for the six-month subgroup (n=22). years post-conference (n=55 children), those placed in out-of-home care. For 4 of these Unfortunately information was not available one to two years post-conference (n=60 chil- children, the move to a non-relative out-of- regarding the length of time between the dren), and those six months to one year post- home care was identified by the family as FGC and case closure. The discussion sec- conference (n=23 children). Except for the their secondary plan. For 7 of the children, tion addresses concerns with relying solely case closure question, findings were virtually their placement in non-relative care was not upon the case closure status as an indicator the same across the 3 subgroups. part of the family plan. The remaining 3 of of well-being.

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 6 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM

DISCUSSION groups or that certain ethnic groups be- Most importantly, this evaluation found that Descriptive findings of this project echo ing more inclined to accept an FGC re- placements remained stable over time with those of earlier Washington State studies, ferral. A greater understanding in these only a few children returning unplanned showing a preponderance of referrals origi- areas is important to assist facilitators in into out-of-home care, and even fewer chil- nating from CWS, high rates of families of their efforts to assure that practice efforts dren being re-referred even two years after color participating in a FGC, and high lev- continue to be culturally responsive. Re- the conference. These positive long-term els of family participation at the conferences call that FGC originated from the Maoris' findings become even more significant when (Vesneski, 1998). Interpretations of these efforts to make the New Zealand child wel- we consider the long-term outcomes for findings were discussed at the facilitator team fare system more culturally responsive. children in Washington State's child welfare meeting and future evaluation questions were system. Although the overall state or regional statistics are not a strict comparison group, identified. For the families within this evalu- This evaluation found ation, FGC appears to be an effective inter- these statistics help to contextualize the find- vention to support families in identifying that placements remained ings from this evaluation. The re-referral their own resources in addition to accessing stable over time with only rate for the families within the evaluation is agency-based supports. Many of the children lower than the rate for the same region. remained or were reunified with their par- a few children returning Overall regional statistics show a 10.1% re- ents, while others were able to remain in the unplanned into out-of- referral rate in 1997(Children's Administra- care of their relatives. tion Services, March, 2000). The percent- home care, and even fewer age found in the Children's Administrative In regard to the small number of referrals children being re- report reflects the "percent of CAN [child originating from CPS, possible interpreta- abuse and neglect] victims who have another tions discussed were: (1) this was a reflection referred even two years substantiated report within 12 months." The of families being less inclined to accept an after the conference. report furthermore states that "the total FGC earlier on in their involvement with the number of referrals is low due to partial child welfare system, (2) a lack of awareness The high level of involvement of family implementation" and consequently the "es- of CPS social workers regarding FGC and timated rates may change substantially in the consequently low referral rates, or (3) an members at the FGC reflects the model's focus on the importance of family mobili- future" (emphasis added). The re-referral effect of possible system disincentives that rate in our study was not limited to a 12 preclude a FGC referral. Dialogue has al- zation and family involvement in the deci- sion-making process. The high rate of pa- month period, but instead counted any sub- ready begun within a CPS unit to gain a stantiated allegation since the date of the greater understanding of the low referral ternal involvement for the families within this study contrasts sharply with previous conference, which could have been as long rates from the perspective of CPS workers. as two and a half years after the conference. This understanding will help identify effec- studies showing very few fathers being in- volved in case-planning, even for family There were substantiated CPS referrals on tive strategies to offer FGCs to families early only two of the 55 children (4%) who were on in their involvement with the child wel- preservation services (O'Donnell, 1999). Consistent with the high rate of paternal in- the focus of the FGC over 2 years prior to fare system. If the outcomes from this evalu- the study. ation serve as an indicator of the effective- volvement at the conference, the rate of ness of FGC for children in general, offer- placement with fathers and/or paternal rela- Although there was a low rate of closed cases ing a FGC to families may help reduce the tives was also high. In regards to the low rate within this evaluation, these rates coincide approximately 31% of the children placed in of children present at the conference, the with the findings of an informal study con- out-of-home care for at least 60 days who current Washington State practice guideline ducted at a Washington State DCFS office remain in care two to three years later (Wash- recommends children over the age of 12 be (Caughey, 1999). In this study, it was found ington Permanency Summit, 1999). At the invited to attend the conference. This guide- that even with the use of innovative and in- facilitator team meeting, members also rec- line differs from New Zealand's where chil- tensive services, it took on average 2 years ommended that even if a family declines a dren younger than 12 are often encouraged for a CWS case to close. FGC initially, there should be repeated of- to attend the FGC. For example, in New The use of case status (open vs. closed) as an fers to have a FGC throughout their involve- Zealand one study found that children were indicator of the effectiveness of FGC should ment with the child welfare system. present at 79% of the conferences (Paterson be used cautiously and in conjunction with and Harvey, as cited in (Robertson, 1996a). other outcome measures. The case closure In terms of the high rate of families of Recently at the FGC team meetings, this color participating in FGCs, it was unclear results are confounded by the fact that guard- guideline for child participation has been ianship cases, even when intended to be per- whether this may be a result of social work- revisited. ers selectively referring certain ethnic manent arrangements, are considered

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 7 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM

"open" in the state of Washington. Guard- HEARING THE FAMILY'S VOICE HEARING SOCIAL WORKER VOICES ianship with relatives was one of the most Although the study results showed positive To improve the FGC process, gathering so- frequently achieved outcomes within this outcomes, we need to enrich our under- cial worker input is also critical. Social work- study. The percentage of cases listed as open standing by hearing from the families them- ers currently represent the gatekeepers be- may also reflect how quickly the legal or child selves. Numbers alone can be misleading tween the families involved in the child wel- welfare system works, rather than the when there is insufficient contextual infor- fare system and the FGC facilitators. In re- progress of the family and the well-being of mation. In this study the collaborative work gards to encouraging social workers to make the child. In other words, the family and the between the FGC facilitators and the Uni- referrals, creating a space for social workers social worker may have completed everything, versity allowed the stories behind the num- to share their FGC experiences, both the but the case remains open waiting for the fi- bers to emerge. For example, for one youth challenges and successes, will help to assure nalization of paperwork in court. Future the plan was listed as "not achieved" because the process meets their needs. Listening to evaluation work should take a closer look as the intended relative guardianship failed. the social workers will help inform future to why a case is still open to determine The facilitator shared, however, that the directions of FGC implementation as well whether, for example, the case is open be- youth eventually was placed with the grand- as how to better address the needs of the chil- cause there is a pending adoption or due to mother, who attended the FGC. This "un- dren and families in their caseloads. the caregiver relapsing. planned" placement has remained stable. Social worker feedback is also critical to gain FUTURE EVALUATION RECOMMENDATIONS Ideally the grandmother's story could also be a greater understanding of how children and heard. For example, what did she think about While the families in this study appear to have families are doing after the FGC and to learn the FGC process and how did it impact her ways to improve the implementation of FGC. benefited from a FGC, additional studies are decision to care for her grandchild? Hear- needed in order to gain a greater under- Inclusion of social worker voices would help ing from the families can also help to assess to more thoroughly assess child and family standing of the long-term outcomes of whether families perceive FGCs as empow- FGCs. These recommendations reflect some well-being. As we have seen from the case ering and can provide an opportunity to closure results, many of the families still have of the limitations of this evaluation. For ex- learn how to improve the process. ample, a reliance on pre-collected informa- an open case even two years after the con- tion restricted our ability to address the full In a desire to learn from families, the FGC ference. If the case is open, social workers range of questions regarding long-term out- facilitators in partnership with the Univer- likely continue to be involved in varying de- comes. Additional work is also needed to sity have created a toll-free survey line for grees with the family. Through this involve- understand whether the positive findings families to self-refer for an interview. In the ment, social workers could provide qualita- from this evaluation are reflective of other interview family members are asked if they tive information on how the children are families' experiences with FGC. would participate in an FGC if they had to faring in their placements. If social workers do it all over again. A mother responded that are providing the required and necessary Recommendations for future evaluation ef- she would as "they are really helpful and use- follow-up support, they could provide long- forts were identified as a result of the find- ful, not just to me, but to my whole family. term information that supplements the more ings from this evaluation and extensive dis- My family learned a lot about me, especially quantitative indicators of well-being tracked cussion with the FGC facilitator team. The how to help me where I need it the most." In by the FGC facilitators. suggestions were generally aimed at identi- another response, a paternal grandmother fying direct strategies to improve the FGC commented that she would "because every- process or ways to expand its use within one was there who needed to be there, and The extended family DCFS. Some of the recommendations ad- stuff got said that needed to be said and offered a tremendous dressed how to improve our ability to mea- people were there to try and rectify it." A sure and track outcome information across maternal grandfather commented that "we amount of support that the state of Washington. Currently the FGC had more information then ever included placement facilitator team is actively working to create before...[and] it allowed the mother a chance options, respite care, and a template of key outcomes and descriptive to speak for herself. She is so used to having variables. This template would allow for ag- everyone tell her what to do, but here she financial assistance. gregation of the data across the different of- really took her issues into her own fices allowing future evaluation work to ex- hands...on a scale from 1-10, it was a 10." amine a larger and more diverse sample. To An additional area of inquiry for social work- Finally a maternal stated, "it was awe- ers is to explore the impact of FGC on how more fully understand the outcomes, sug- some for us, the family, to see my sister be gestions were also made that require the col- workers and the agency approach families. the lead of her situation and feel like she was In Pennell and Burford's (Pennell & lection of qualitative data gathered from both in control." the families and the social workers. Burford, 2000) study, they noted an inter-

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 8 Restorative Practices September 10 2002 www.restorativepractices.org

E FORUM esting secondary impact of their project, one pilot projects in and . Simmonds, J., Bull, H., & Martyn, H. that lasted after the family group decision- Sheffield: University of Sheffield Part- (1998). Family Group Conferences in making project concluded. Workers reported nership Research Programme. Greenwich Social Service. London: that they were more likely to engage extended Hudson, J., Morris, A., Maxwell, G., & Goldsmith's College. family in a variety of ways to respond to the Galaway, B. (Eds.). (1996). Family Swain, P. (1993). Safe in our hands — the child's well-being. This approach was ex- Group Conferences. Leichhardt, New evaluation report of the family deci- tended to families regardless of whether they South Wales, Australia: The Federation sion-making project. Melbourne, Aus- were involved with a formal family group de- Press. tralia: Mission of St. James and St. cision making process. The facilitator team Johns. has shared anecdotal evidence that suggest Immarigeon, R. (1996). Family Group similar stories with the social workers here Conferences in Canada and the United in Washington State. An exploration of States: An Overview. In J. Hudson, A. whether FGCs impact overall agency culture Morris, G. Maxwell, & B. Galaway with regard to family engagement in care (Eds.), Family Group Conferences: represents yet another area for future evalu- Perspectives on Policy and Practice (pp. ation work. 167-179). Monsey, N.R.: Willow Tree Press, Inc. CONCLUSION Lupton, C., & Nixon, P. (1999). Empow- FGC appears to be an effective way to move ering Practice? A critical appraisal of beyond the metaphor of the pendulum. For the family group conference approach. the children within this evaluation, the im- Bristol, United Kingdom: The Policy mediate and long-term outcomes suggest the Press. children were protected and the family unit Lupton, C., & Sheppard, C. (1999). Fam- was honored. Both the maternal and pater- ily outcomes: following through on nal sides of the family participated at the family group conferences. Portsmouth, conference and with case planning. The ex- U.K.: Social Services and Information tended family offered a tremendous amount Unit, University of Portsmouth. of support that included placement options, respite care, and financial assistance. This O'Donnell, J. M. (1999). Involvement of support reinforces the belief that extended African American fathers in kinship families must be brought into the decision foster care services. Social Work, 44(5), making process for families involved in the 428-441. child welfare system. Additional evaluation Pennell, J., & Burford, G. (2000). Fam- work, including the recommendations out- ily Group Decision Making: Protecting lined within this paper, will further our un- Children and Women. Child Welfare, derstanding of the effectiveness of FGC and LXXIX(2, March-April), 131-158. help to identify ways to improve the process Robertson, J. (1996a). Research on Fam- for both the families and social workers. ily Group Conferences in Child Wel- fare in New Zealand. In J. Hudson, A. REFERENCES Morris, G. Maxwell, & B. Galaway Bodonyi, J, Kemp, S. "Characteristics of (Eds.), Family Group Conferences: Legally Free Children and Recommen- Perspectives on Practice and Policy (pp. dations for Permanency in Washington 49-64). Monsey, New York: Willow State", a Report for the Families for Tree Press. Kids Initiative, A Public-Private-Tribal Partnership, August 23, 1996 (rev.) Robertson, J. (1996b). Research on Fam- ily Group Conferences in Child Wel- Caughey, B. (1999). conversation around fare in New Zealand. In J. Hudson, A. case tracking data. Morris, & G. Maxwell (Eds.), Family Crowe, G., & Marsh, P. (1997). Family Group Conferences (pp. 49-64). Group Conferences, partnership, and Leichhardt, New South Wales, Austra- child welfare: a research report on four lia: The Federation Press.

INTERNATIONAL INSTITUTE FOR RESTORATIVE PRACTICES 9