Accomplishing family reunification for children in care: An Australian study☆

Abstract Whilst child protection systems are concerned with removal of children from their families in the interests of safety, the capacity of child welfare systems to return children safely to their families of origin is of central importance. The multidimensional standardised assessment tool, the North Carolina Family Assessment Scale—Reunification (NCFAS-R) was used by practitioners to assess family strengths and needs in case planning and reunification decision making. The current paper examined (1) whether NCFAS-R domain ratings at intake and closure differ by characteristics of parents and children; and (2) whether reunification is predicted by NCFAS-R score at closure. The study sample consists of 145 children aged 0–12 years from 84 families, who presented at Barnardos temporary care services in two metropolitan areas in Australia. This excludes children who had missing values on NCFAS-R or reunification outcome. Participants continuously entered the study over the four year study period, the study window being 18 months since intake. Ordinary least squared (OLS) regression was used to examine whether NCFAS-R scores at intake and closure were predicted by demographic variables, primary reason in care, and placement circumstance. To examine the relationship between NCFAS-R scores at closure and reunification outcome, a logistic regression model was used. At intake, the average score was highest for the Child Well-Being domain and lowest for the Parental Capabilities domain. NCFAS-R scores were increased at closure on all domains, with the biggest improvement on the domains of Family Safety and Child Well-Being. At intake, NCFAS-R scores did not differ significantly by independent variables examined except for the Child Well-Being domain. Children who were placed with their siblings displayed 0.45 points higher scores on the Child Well-Being domain. At closure, NCFAS-R scores differed significantly by some family variables and a placement variable. In general, mothers being 25 years or younger, mothers having Year 11 or a higher level of education, or children being placed with their siblings were significantly associated with higher scores on various NCFAS-R domains at closure. Overall NCFAS-R scores at closure significantly predicted reunification with parents or kin. One unit increase in overall NCFAS-R score at closure increased the odds of reunification by a factor of 8.39. Findings contribute to an evolving evidence base on decision making and facilitating reunification outcomes for children and families.


Introduction
Reunification is at the centre of meaningful child welfare practice. Whilst child protection systems are concerned with removal of children from their families in the interests of safety, the capacity of child welfare systems to return children safely to their families of origin is also considered to be of central importance. Yet reunification has tended to remain a largely invisible area of work (Farmer, Sturgess, O'Neill, & Wijedasa, 2011;Pine, Spath, & Gosteli, 2005).
Child welfare services have emphasised supportive work with biological families to prevent abuse and neglect and removal of children into protective care. When placement in care is needed the goal is to reduce the length of separation between parent and child, and to maximise the prospects of reunification of children with their parents or kin whenever it is safe to do so (Berrick, 2009). Apart from the economic costs of maintaining children in care, research has highlighted the undesirable consequences for children of remaining in care for long periods. Extended periods of time in care can lead to loss of family connections and a sense of identity, and difficulties in transitioning out of care (Pecora et al., 2005). For those experiencing multiple placements there is evidence of later difficulties in forming attachments with adults and of developing long term emotional and behavioural problems (Stovall-McClough & Dozier, 2004).
The demands placed on the system by the volume of children entering care imposes constraints on its capacity to maintain effective case planning, including reunification, and high standards of care. In 2011 there were 37,648 Australian children aged younger than 18 years old in out-of-home care (7.3 per 1000 children) a figure which has consistently risen every year in the last decade, and by 33% since 2007 according to the Australian Institute of Health and Welfare (2012). Just under half (42%) of the children entering care in 2010-2011 were aged less than 5 (4879). Of these the majority (93%) were in home-based care with roughly equal numbers in foster care (45%) and relative/kinship care (46%) with only one in twenty children living in residential care (Australian Institute of Health and Welfare, 2012). Indigenous children in out-of-home care were overrepresented in all States and Territories, ranging from the lowest rate reported in the Northern Territory (18.2 per 1000 children) to the highest rate reported in New South Wales (80.6 per 1000) (Australian Institute of Health and Welfare, 2011).
Family reunification knowledge and research is limited, particularly in Australia, despite documented evidence that most children placed in protective care are eventually reunited with their birth parents (Child Welfare Information Gateway, 2011;Fernandez & Delfabbro, 2010). Given this context, the research reported in this paper explores how knowledge of family characteristics, needs and strengths can contribute to reunification decision making and practice in child welfare, and address an important Australian and international knowledge gap.

The Australian context
The Australian context of statutory child protection is the responsibility of State and Territory governments, and as a result, rather than a single national system, there exists eight different child protection systems, with broadly similar processes but each with its own legislative framework, policies, procedures and practices (Bromfield & Higgins, 2005). For instance placement in Out-of-Home Care (OOHC) can be by court orders issued under the New South Wales Children and Young Persons (NSW Care and Protection) Act 1998 or under voluntary request/agreement. At 30 June 2011, 85.5% of the 16,740 children in out-of-home care in New South Wales were on a care and protection order, which is roughly similar to the national pattern for order status (Australian Institute of Health and Welfare, 2012). The NSW Care and Protection Act 1998 requires priority to be given to the child's right to be raised in the biological family and prevent placement, and if separation becomes necessary, planned return of the child as soon as possible to the family.
As of October 2011, 1990 children in New South Wales were placed in non-government organisation statutory placements mostly in general foster care, intensive foster care, or residential care, with small numbers in relative/kinship care group or semi-independent living arrangements (Ministerial Advisory Group, 2011). The agency delivering the programme which is the research site for the present study, Barnardos Australia, is a major charitable child welfare services provider and provider of statutory out of home care placements in New South Wales (NSW) and the Australian Capital Territory (ACT).

Previous research on reunification
Previous research studies have been undertaken internationally to isolate the variables associated with reunification outcomes. The timing of reunification is the focus of several studies. Trends in various studies suggest that many children are reunified rapidly and that the likelihood of return declines after six months. In an American study, Wells and Guo (1999) found that 36% of children were reunified within 24 months of being placed in care whilst Taussig, Clyman, and Landsverk (2001) assert that between 50 and 75% of children placed in care eventually reunify. Wulczyn (2004) reporting from the Multistate Foster Care Data Archive notes that overall in the US the first year a child is in foster care the probability of reunification is 28%. This probability drops to 16% over the following year and as time goes on the probability of reunification declines. Key studies by Wade, Biehal, Farrelly, and Sinclair (2011), Fernandez and Lee (2011), McSherry, Weatherall, Larking, Malet, and Kelly (2010, Connell, Katz, Saunders, & Tebes, 2006), Delfabbro, Barber, and Cooper (2003), Fernandez (1999), Bullock, Gooch, and Little (1998), Barth, Courtney, Berrick, and Albert (1994), and Fanshel and Shinn (1978) report similar reunification patterns.
Predictor variables most commonly analysed in outcome studies on reunification are, age of the child, gender, ethnicity, reasons for placement and placement type. A child's age is associated with patterns of return. The likelihood of speedy return is lower for those who enter as infants (Leathers, Falconnier, & Spielfogel, 2010;Sinclair, Baker, Lee, & Gibbs, 2007). Very young children returned home at a slower rate, whilst adolescents were more likely to experience rapid return (Fernandez & Lee, 2011). Characteristics of the families and their children who are to be reunified, or elements in the family's environment, also have been found to influence reunification. Children with health problems and/or disabilities were found to return at lower rates (Barth et al., 1994). A large scale study found that children displaying behaviour or emotional problems as indicated by CBCL scores found that they were 50% less likely to be reunified (Landsverk, Davis, Ganger, Newton, & Johnson, 1996). In Australia, Aboriginal and Torres Strait Islander children were found to be significantly less likely to reunify (Fernandez & Delfabbro, 2010). Other family characteristics that have a negative impact on reunification are poverty and environmental stress, inadequate or unstable housing, single parent status and financial difficulties. Family disadvantage was a robust predictor of delayed or non-reunification in many studies (Fernandez, 1996;Jones, 1998;Kortenkamp, Geen, & Stagner, 2004). Children from single parent families were three times less likely to return Wells & Guo, 1999).
The proportion of children restored to parents is lower for children whose families experience complex problems. Parent profiles associated with reduced probability of reunion included mental illness, emotional problems, substance abuse and domestic violence (Fernandez & Lee, 2011;Goerge, 1990;Jones, 1998;Marsh, Ryan, Choi, & Testa, 2006). Multiple and co-occurring problems such as lack of supervision, poor parenting skills, domestic violence, and mental health amongst birth mothers tend to have a negative effect on the reunification process (Cheng, 2010;Choi & Ryan, 2007;Fernandez & Lee, 2011). For example, in a study examining the speed of reunification with parents, Fernandez and Lee (2011) found that, compared to children with parental health concerns, children with parental substance abuse issues had 86% lower rate of reunification and children from domestic violence situations or other issues had 73% lower rate of reunification with their parents. A comprehensive study by Shaw (2010) concluded that families experiencing parental drug or alcohol use have lower odds of reunification compared to those in which parents do not have any indications of these conditions. Substance abusing mothers who utilised child care services were more likely to achieve reunification in a US study (Choi & Ryan, 2007). However, children in the out of home care system are seldom there for any single reason. Whilst there may be in fact an overtly identified problem such as parental drug use, there is commonly a cluster of co-contributing factors which have led to the child being placed in care.
Research interest in the association between parental visits and reunification outcomes has identified important trends. In their study of 925 children, Davis, Landsverk, Newton, and Ganger (1996) found that visits were the key to discharge from care. When visit plans were developed, the likelihoods of visits were increased; the majority of children who visited with their parents at the level recommended by the courts were reunified with their families (Bullock et al., 1998;Davis et al., 1996;Farmer, 1996). Berry, McCauley, and Lansing (2007) state the most significant predictor of whether a child will be reunified to birth parents is the amount of contact shared by both parties. An evaluation of the evidence on contact is available in Biehal (2006b).
The availability of services to birth families to ameliorate the problems precipitating placement initially or necessitating its continuation is also widely documented in the analysis of reunification outcomes (Biehal, 2006a(Biehal, , 2006bMaluccio, Abramczyk, & Thomlison, 1996;Thoburn, Robinson, & Anderson, 2012;Littell & Schuerman, 1995). For example, a study by Fraser and his colleagues demonstrated significant success from using a 90-day intensive family preservation model with separated families, which achieved a 93 percent reunification rate with the fifty-seven children whose families were involved in the programme (Fraser, Walton, Lewis, Pecora, & Walton, 1996). The use of clear case plans accompanied by professional support for parents and children during the reunification process is stressed. Proactive social work, effective case planning and a high level of social work involvement has been shown to facilitate reunification (Farmer, 1996;Farmer et al., 2011). Previous studies have attributed success to service variables such as the provision of concrete services, the establishment of strong worker-family relationships, and the provision of education and training to parents (Barth et al., 2005;Fraser et al., 1996;Lewis, 1994;Walton, Fraser, Lewis, & Pecora, 1993). Staff and Fein (1994) reporting on another experimental family reunification programme also found that concrete assistance to the families involved promoted reunification. Family engagement and use of services is seen to have a positive effect on reunification (National Clearinghouse on Child Abuse and Neglect Information, 2005). Mental health problems and disabilities are perceived to pose risks for not achieving reunification. Akin (2011) andLandsverk, Burns, Stambaugh, andReutz (2009) argue from their findings that child welfare services must recognise disability and mental health problems as barriers to reunification and ensure access to evidence based health services for children and families. The size and quality of the families' social support network have shown to be positively correlated with reunification (Festinger, 1996;Fraser et al., 1996).
Prior child welfare involvement is considered to impact on prospects of reunification. Connell et al. (2006) found that those children who had experienced two or more removals previously were significantly less likely to be reunified. A larger number of placement moves has also been commonly associated with longer periods in care prior to reunification Fernandez, 1999;Wells & Guo, 1999;Wulczyn, 2004).
The robustness of re-unions and factors that contribute to reunification breakdown are explored in various studies. Wade et al. (2011) have reported high rates of reunification breakdown, with the majority (two-thirds) of all returned children in their study not being at home continuously after a four years follow-up. Due to limited post-reunification support and services, there seems to be a significant amount of children re-entering care once reunification fails or breaks down. The lack of comprehensive assessments and resolution of the problems that first precipitated entry into care are cited as significant factors in re-entry into care (Fraser et al., 1996;McDonald, Bryson, & Poertner, 2006). In a UK follow-up study of foster children, Sinclair, Baker, Wilson, and Gibbs (2005) found that children who returned home experienced a significantly higher re-abuse rate than children who remained in care. The apparent lack of planning and post reunification support is cited (McMurtry & Lie, 1992). Data indicate that 20-40% of reunified children re-enter foster care ) a trend consistent with estimates of a third claimed by Wells and Guo (1999) and the 37.0% claimed by Terling (1999). According to Barth, Weigensberg, Fisher, Febrow, and Green (2008), two main factors are associated with re-entry; Children having higher problem scores on the CBCL, and children being reunified to a household accommodating three or more children post-reunification.
Despite rates of reunification and re-entry being increasingly documented few studies feature more extensive outcome measures including indications of improved family functioning at return. There is a need for locally based contemporary studies to better inform case decision making in the Australian context. The study reported in this paper is distinctive in focusing on reunification processes and outcomes in the context of an Australian temporary family care programme providing crisis care, respite care and bridging care services for families. This paper draws on data from a larger research project that explored the process and outcomes of reunification through a four year prospective longitudinal study using quantitative and qualitative approaches. Specific aims of the research were to: ▪ gain an understanding of pre and post intervention characteristics, circumstances and functioning of families whose children enter temporary foster care ▪ identify child and family characteristics of the studied sample that are associated with reunification ▪ determine the relative influence of case characteristics, service variables on reunification outcomes The current paper is focused on family characteristics and functioning that are associated with reunification. In particular the paper explores whether family needs and strengths predict reunification outcomes.

Study design and sample
Five research sites were studiedthree family centres in Greater Metropolitan Sydney, New South Wales (NSW); one family centre in rural and regional New South Wales; and one family centre in the metropolitan region of Canberra, Australian Capital Territory (ACT). The study was carried out at Barnardos (hereafter referred to as the agency) Temporary Family Care (TFC) programmes which provide integrated assessment, family preservation and out-of-home care services to families with children aged 0-12 years at intake, who are 'in need', or who are at 'significant risk of harm'. TFC programmes developed by Barnardos Australia as part of their continuum of permanency planning operate in areas of NSW and ACT. TFC is seen as the initial point of contact with a target of permanency with the family of origin. Developed more than 25 years ago, the TFC model provides crisis placements and foster placements with carers who are committed to reunification to parents wherever possible. The service model focuses on setting up ongoing help for families through referral to relevant services, to support families to care safely for their children after a time-limited foster care placement. Where it is not safe for a child to return home, caseworkers initiate a plan for long stay foster care or adoption. Parental visiting and parent/foster carer contact are part of the case plans to reduce a child's feelings of separation and loss.
The participants in this study were derived from a consecutive sample of children and families who presented at Barnardos temporary care services in two metropolitan areas. The sample included 168 children from 96 families. All children were entering out of home care for the first time. A minimum period of two weeks in care was the criterion for inclusion in the study. Participants continuously entered the programme throughout the study period, the study window being 18 months since intake. Participants were followed up to restoration or up to the 18 months since entry into the care (whichever is first).
Data were collected through face to face interviews conducted onsite at Barnardos' TFC Centres with 103 caseworkers. Interviews with caseworkers took place within 6 weeks of intake (Intake 1) with a follow up interview later six months, or at the time of restoration depending on which came first (Closure 2). The study explored caseworkers' perspectives on the decisions regarding placement and reunification. Caseworkers responded to questions about the background and functioning of families they worked with, the reasons the child/ren came into care, and the case plans instigated as part of the restoration process. Rich descriptions of practitioners', carers' and parents' experiences during the reunification process, and the meanings ascribed to them were captured through the qualitative data. A thematic analysis was used to identify emerging themes (Alston and Bowles, 2003) which are reported more comprehensively in Fernandez (2012). In addition, a standardised assessment tool, the North Carolina Family Assessment Scale-Reunification (NCFAS-R) (Reed-Ashcroft, Raymond, & Fraser, 2001) was completed by caseworkers to capture family functioning at entry or Intake and at Closure of the case or at restoration.
The data analysed from caseworker interviews provide a profile of the families and children and pathways into care. Above half (54.5%) of mothers were aged between 26 and 35 years. The vast majority (86.6%) were receiving benefits as their primary source of income. A mere 3.7% of families owned their homes, 4.9% were homeless at the time of entry. The bulk of families (70.5%) were living in Public Housing or Community accommodation. In terms of family composition, 6% of families had 8 or more children in the family, 24% had between 5 and 7, 35% had either 3 or 4 and 35% had 1 to 2 children. It was not uncommon for children to enter care as sibling groups. Seventy five percent of families had 2 or more children in care and 66% of children were placed together in the same placement. The primary source of referral was the Statutory Department (76%). It is worth noting that 15% referred themselves voluntarily.
As part of a larger research project investigating the process and outcomes of reunification, the analysis reported in this paper examined the changes in NCFAS-R scores between intake and closure to establish 1) whether NCFAS-R scores at closure differs by family characteristics (i.e., mother's age, mother's education) and placement situation (i.e., reason in care, siblings in the same placement), and 2) whether reunification is predicted by NCFAS-R scores. The sample for the current analysis comprised of 145 children from 84 families after the exclusion of 23 cases with missing values on NCFAS-R or reunification status.

Variables
Outcome variables were reunification status and the North Carolina Family Assessment Scale-Reunification (Reed-Ashcroft et al., 2001). Reunification was defined as being restored to parents or kin. During the study period 53% of children were restored to either their birth parents or kin. The North Carolina Family Assessment Scale-Reunification (NCFAS-R) is a modified version of the North Carolina Family Assessment Scale to be used in working with reunification cases (Reed-Ashcroft et al., 2001). The NCFAS is an instrument to assess family functioning and social environment and measure outcomes. The NCFAS has been considered as one of best instruments to be used by practitioners to assess areas needing service (Johnson et al., 2008;Kirk, Kim, & Griffith, 2008). The NCFAS-R has internal consistency reliability ranging from .71 to .94 and its construct validity and criterion validity have been verified (Kirk et al., 2008;Reed-Ashcroft et al., 2001). The NCFAS-R covers seven domains: Environment, Parental Capabilities, Family Interactions, Family Safety, Child Well-Being, Caregiver/Child Ambivalence, and Readiness for Reunification. Each domain has 3 to 10 items (refer to Table 1 for details). Each item was rated on a six-point scale (serious problem, moderate problem, mild problem, baseline/ adequate, mild strength, and clear strength). Raw scores range from 2 (clear strength) to −3 (serious problem). For the current analysis, scores are converted to 5 (clear strength) to 0 (serious problem).
Other variables included child's age, mother's age, mother's education level, primary reason in care, and sibling in the same placement. Child's age at the entry into care was measured in years. Dummy variables were created for mother's age and education level. Mother's age was grouped to three categories: 25 years or younger, 26 or older (reference group), and unknown. Mother's education level was categorised as Year 10 or less (reference group), Year 11 or more, and unknown. The primary reason in care was grouped into two categories: parental health issues coded as '1' and other issues coded as '0'. Other issues included child abuse and neglect, parental substance abuse, domestic violence, parenting difficulties, parental imprisonment and other. This decision to divide primary reasons into these categories was based on findings of the authors' previous analysis (Fernandez & Lee, 2011). In this study using an event history analysis model, it was found that parental health issues significantly predicted the speed of reunification. Siblings in the same placement was a dummy variable with '1' indicating that siblings are placed together.

Data analysis
Sample characteristics were examined using univariate statistics. To identify patterns of NCFAS-R scores at intake and closure, we examined descriptive statistics and correlations were examined. Ordinary least squared (OLS) regression was used to examine whether NCFAS-R scores at intake and closure were predicted by demographic variables, primary reason in care, and placement circumstance. To focus on the change of NCFAS-R score between intake and closure, NCFAS-R score at intake was controlled for the analyses of NCFAS-R score at closure. To examine the relationship between NCFAS-R score at closure and reunification outcome, logistic regression model was used because reunification outcome is a dichotomous variable. After controlling for the effects of the other variables, multivariate logistic regression estimates the effect of NCFAS-R score at closure on the likelihood of being reunified with their parents or kin by Overall well-being, child's mental health, child's behaviour, school performance, relationship with caregivers, relationship with siblings, relationship with peers, motivation/cooperation to maintain the family Ambivalence Overall caregiver/child ambivalence, parent/caregiver ambivalence towards child, child ambivalence towards caregiver Readiness for Reunification Overall readiness for reunification, resolution of significant CPS risk factors, completion of case service plans, resolution of legal issues, parent/caregiver understanding of child's treatment needs, established back-up supports or service plans In the current study, Readiness for Reunification was not used to assess social environment at intake.
presenting the odds ratios. The multiple imputation procedure was employed to reduce possible bias due to missing values and to make valid statistical inferences (Fichman & Cummings, 2003). Note, however, that outcome variables were not imputed. All analyses were performed using IBM SPSS Statistics for Windows, Version 20.0 (IBM Corp, 2011).

Descriptive statistics
Among the study sample, about 52% of children were reunified with their parents or kin. Children's age at entry ranged from 3 days to 15 years-old and the mean age was about 5.91 (SD = 4.05). About 12% of mothers were 25 years or younger and about 65% of mothers were older. About 41% of mothers finished Year 10 or less and about 9% of mothers completed Year 11 or more. About 22% of children entered care due to parental health issues whereas the rest of children entered care due to other issues such as child abuse and neglect (28.3%), domestic violence (12.4%), or parental substance abuse (20.7%). Almost 56% of children were placed with their siblings. Table 2 displays descriptive statistics of the study sample.

Changes of NCFAS-R scores between intake and closure
First, a brief overview of patterns of family functioning assessed by workers at intake using the NCFAS-R follows. Substantial difficulties were experienced by families across a number of the NCFAS-R domains: in terms of overall environment, nearly half (49.0%) of families were experiencing moderate and serious problems, which included problems with the learning environment, financial management, and food and nutrition. Only about 12.5% of families displayed strengths in their overall environments. In terms of overall parental capabilities, majority of families (70.8%) were seen to be experiencing moderate to serious problems. The subscale of supervision of children reflecting the greatest concern, where two-thirds (66.4%) of families were experiencing moderate to serious problems at intake. In nearly half (47.7%) of these cases substance abuse was considered a serious problem affecting parenting. In terms of overall family interactions, nearly half (45.5%) of families were experiencing moderate to serious problems. The domain of Family Safety showed that more than half (55.9%) of families were experiencing moderate to serious problems, including neglect (65.7%), emotional abuse (50.3%), and domestic violence (60.4%). Only a small minority of families (6.4%) were considered to possess strengths in regard to overall family safety.
For the category of 'overall child well-being', problems of moderate to serious nature were identified in 30.4% of families. Several sub scales for the domain of Children's Well-Being showed that despite overall concerns with environment, parental capabilities, family interactions and safety, families still displayed a higher level of strengths evidencing perhaps a capacity for resilience in the children. Only a small sub-group (15.6%) of families displayed any moderate to serious problems in co-operation and motivation to maintain the family and most families (83.0%) were rated as being either on the baseline or having clear to mild strengths. Despite the multiple stressors observed on parent-child relationships, 56.6% of families were rated as being at baseline or above at intake on the relationship with parent(s)/caregivers subscale. For overall caregiver/child ambivalence, 16.6% displayed moderate to serious problems. However the subscales for this domain varied by the direction of ambivalencewhereas only 15.4% of families experienced moderate to serious problems with child ambivalence to the parent/caregiver, double the rate (32.6%) were considered to have moderate to serious problems with parent/caregiver ambivalence expressed towards the child. The latter trend is important as the qualitative interview data showed that caseworkers viewed perceived parental empathy and engagement with the child as a key factor in predicting the likely success of reunification (and hence in assessing the suitability of reunification). This view of frontline workers is consistent with previous research by Hess, Folaron, and Jefferson (1992) that a significant impediment to implementing successful reunification is parental ambivalence. Overall, about half of families (50.3%) displayed moderate to serious problems in their readiness for reunification whereas only 17.3% of families showed strength in this domain.
In terms of the seven family functioning domains of NCFAS-R at closure, substantial changes in the level of need and/or risk of harm were witnessed across a number of these domains. Summarising the observed trends at the level of each of the overall domains the following patterns were noted: Overall environment most notably improved in the moderate to serious problem rating for close to half of families (49.0%) decreasing to 29.8% at Closure. Overall parental capabilities rated in the moderate to serious problem category decreased from 70.8% of families to 39.3% at closure. Overall perceived strengths in family interactions increased positively. Families rated as having clear and mild strength increased from 18.6% to 26.2%; whereas an even greater increase was observed in families rated as baseline from 6.9% to 22.1%. Overall family safety showed considerable improvement most notably in those considered having "clear to mild strengths" which increased from 6.8% of families at intake to 26.2% at closure. Those experiencing moderate or serious problems decreased from more than half (55.9%) of cases to a third (34.5%) by closure.
The Child Well-Being domain reflected significant improvements, moderate to serious problems declining from 30.6% to 8.4%. In addition the number of families displaying strengths increased from 19.5% to 51.8%. Overall caregiver/child ambivalence findings were mixed depending on the level of severity of the problem. The number of families displaying strengths increased considerably from 16.5 % at Intake to 33.4% at closure. Whilst 44.1% families displayed mild problems before entering TFC, this had decreased to 15.9% of families by closure, whereas those considered to have moderate to serious problems showed less improvement, decreasing from 16.6% to 13.1% of all families. Overall readiness for reunification showed improvements in families presenting strengths increasing from 17.3% at intake to 38.2% post-intervention, whereas those families experiencing moderate to serious problems showed a smaller change (a decrease in the proportion of families affected from 50.3% to 40.3%). Taking the latter group together with the 9.7% of families still experiencing mild problems post-intervention, half of the sample (50.0%) still displayed problems that in their caseworkers' estimation impaired their readiness for reunification. This is consistent with the rate of reunification (52.4%) observed in the study. The number of cases not restored (47.6%) most likely reflects the persistence of problems reported in this domain i.e. half of families were still experiencing problems with regard to their preparedness to reunify. The reverse also holds true where families had achieved either the baseline functioning or at the level of a strength, in nearly all cases they were reunified. Table 3 displays correlations among NCFAS-R domain scores. As we expected, domains of NCFAS-R were highly correlated, with the Pearson's correlation coefficients ranging from .19 to .68 at intake and from .25 to .81. For the main analyses, composite scores of NCFAS-R were examined. At intake, the average score was highest for the Child Well-Being domain (M = 2.77, SD = 1.01) and lowest for the Parental Capabilities domain (M = 1.65, SD = 0.86) as shown in Table 3. For all domains, NCFAS-R scores increased at closure with the strongest improvement reflected in the domains of Family Safety (0.72) and Child Well-Being (0.69). At closure, the average score was also highest for the Child Well-Being domain (M = 3.43, SD = 0.90) and lowest for the Parental Capabilities domain (M = 2.24, SD = 1.13). See Table 4 and Fig. 1 for details.
The analysis also examined whether NCFAS-R scores were predicted by demographic variables such as, primary reason in care, and placement circumstance. Details are displayed in Table 5. At intake, NCFAS-R scores did not differ significantly by independent variables examined except for the Child Well-Being domain. Children who were placed with their siblings displayed 0.45 points higher scores on the Child Well-Being domain (p b .05).
At closure, mother's age, mother's education level, and placement with siblings significantly predicted some domains of NCFAS-R, even after controlling for NCFAS-R scores at intake. Compared to their counterparts, mothers who were 25 years or younger had significant-

Predicting reunification using NCFAS-R
Due to high correlations among domains of NCFAS-R, overall NCFAS-R score was used for logistic analysis to examine the relationship between NCFAS-R score at closure and reunification outcome. Holding all other variables constant, overall NCFAS-R score at closure significantly predicted reunification with parents or kin. One unit increase in NCFAS-R score at closure increased the odds of reunification by a factor of 8.39. See Table 6 for details.
In addition, we examined whether groups received different services during care because NCFAS-R score at closure was a significant predictor of reunification and changes of NCFAS-R score differed by groups. Given the limited size of study sample, separate chi-square tests were conducted. Other things being equal, younger mothers were less likely to utilise educational services; more educated mothers were more likely to utilise educational services. As expected, when children were placed in care due to parental health issues, these families were more likely to receive health related services. Families of children who were not placed with their siblings were more likely to use educational and financial services. See Table 7 for further details.

Discussion and implications
Decisions about interventions and services to be provided to children and families should be grounded in comprehensive assessments (Johnson et al., 2008;Rycus & Hughes, 2008). A review of practices  that are helpful in reunifying families indicate an emphasis on comprehensive assessment of the strengths and needs of children and families, building on strengths and addressing needs through responsive service delivery (Child Welfare Information Gateway, 2011).
Initial individualised needs assessments are critical to the formulation and implementation of case plans that lead to reunification. The literature is also emphatic about needs and safety assessments to be carried out prior to reunification to minimise risk and harm to children and re-entries to care. This requires comprehensive assessment that goes beyond focusing on incidents precipitating entry to placement   to analyse family functioning, parents' individual health and emotional functioning, and community and social environment. The use of standardised tools in assessment of base line needs and readiness for reunification is an evolving area of research and practice that offers potential for improving case planning and decision making in child welfare (Corcoran, 1997). Actuarially based risk assessment scales are used increasingly in child welfare systems to support decision making about children's risk of harm. For a further discussion relating to risk assessment see Gambrill and Shlonsky (2001) and Johnson et al. (2008). Such assessments benefit from the use of standardised instruments such as the NCFAS-R to identify needs, strengths and areas of risk at intake and return. Whilst the goal for most children in care is reunification with families, premature or inappropriate decisions to reunify can compromise children's safety. It is expected children will be returned home when threats to their safety and well-being are significantly reduced, family protective capacities and strengths are enhanced, and child and family problems are sufficiently ameliorated. Agencies are increasingly adopting structured protocols to identify and respond to needs of families. The use of such standardised protocols facilitates consistency in assessment criteria and in depth assessments of issues of safety of the child and parenting capabilities. In this research data on reasons for protective care nominated by caseworkers were complemented by their completion of the standardised measure of family need and functioning, the NCFAS-R. Family functioning measured against the NCFAS-R at the time of entering care reflected high levels of need and risk. Substantial difficulties were experienced by families across a number of NCFAS-R domains. Prior to deciding to reunify caseworkers completed a reassessment of risk through assessing the family's functioning on the NCFAS-R. Between intake and closure most families recorded improved family functioning on all domains with Family Safety and Child Well-Being showing the most improvement. The trends in scores illustrate the amount of measureable change achieved during the service period from intake to closure ratings. Results indicated that NCFAS-R assessment at closure predicted reunification outcome. The odds of reunification increased proportionally with incremental improvements in strength ratings on all domains. In terms of relationships between demographic variables and NCFAS-R scores the strongest gains were demonstrated by younger mothers and mothers with higher levels of education. Thus NCFAS-R appears to be a promising assessment instrument for use in reunification decision making with its strength based orientation enhancing understanding of family needs and assets to guide decision making and therapeutic strategies. Overall NCFAS-R data provided a useful multidimensional measure of needs and change in this cohort of families. The seven domains provide helpful scope for the rating approach allowed for examination of strength acquisition as well as problem reduction. The potential of the instrument was further demonstrated in previous analyses. In Fernandez and Lee (2011), a risk typology based on caseworker ratings of the NCFAS-R was developed through latent profile analysis and the typology was used to predict the speed of reunification in an event history analysis. The results similarly showed that, compared to children of families with low risks and high strength ratings, children with high risks had 73% lower speed of reunification with their parents.
Whilst capturing fidelity measures was outside the scope of this study, this is perhaps indicative of good fidelity in implementation in relation to the caseworkers' practices. Consistent with the programme's philosophy, this evidences that once caseworkers have reached a clinical judgement that family functioning has been sufficiently enhanced and pre-existing problems and support needs have been addressed they have been vigorously following through on plans to enable children to return home, and are thus ensuring children are not kept in care placements any longer than family circumstances necessitate.
Domains of NCFAS-R represent critical aspects of family functioning and have been the foci of interventions. Complementing the quantitative assessments derived from NCFAS-R ratings in an analysis of qualitative data from the larger research project it is still evident, that caseworkers also based their decisions about reunification on their perceptions of the mother/parent's attitude to the children. These perceptions include the parent's level of empathy for the child, their level of insight into their child's needs and their ability to engage effectively with their child (Fernandez, 2012). Observed improvements in the attachment bonds between parent and child throughout the contact visits was also factored into their decisions. The emphasis on attachment and engagement is consistent with prior research by Carlson, Smith, Matto, and Eversman (2008) where maternal readiness was identified as a key factor in successful reunification. Similarly, this also reinforces previous research findings highlighting the importance of contact visits in facilitating reunification by strengthening parental attachment bonds (Berry et al., 2007).
Certain characteristics pertaining to contact, and the quality of relationship between carer and parent, caseworker and carer, and caseworker and parent appear instrumental in reunification outcomes. Previous research cites evidence which highlights the importance of caseworkers in facilitating reunification through supporting on-going contact with children and encouraging parental involvement in joint planning. Contact has been identified in prior research as a central factor in reunification increasing the likelihood of reunification and enhancing the potential process of reintegrating the child into the family (Wulczyn, 2004). Maintenance of contact between child and birth families and its subsequent impact on reunification outcomes has been a fertile ground for research. (Berry et al., 2007;Bullock et al., 1998;Cleaver, 2000;Davis et al., 1996;Delfabbro et al., 2003;Fernandez, 1996;Leathers, 2002). Davis et al. (1996) demonstrated in their study of 925 children, that contact visits were fundamental to reunification. A majority of children who had contact with their parents as suggested by the courts were successfully reunified. The impact and outcomes of contact between foster children has been contentious with pro and anti arguments grounded in theory and research. The issues are less complex where the intent is return and continued parental contact is essential (Kelly, 2000). In this study, contact was frequent and was in general viewed positively by all three parties, with caseworkers observing that contact between parents and child often enhanced the bond, without the added stress that caring for the child everyday contributed. During care episodes 97% of children in this study had contact with a parent or sibling: 60% had weekly contact and 28% fortnightly. Most researchers endorse that visits must be planned, addressing issues of location, preparation, support, perspectives of children, parents, carers and social workers (Leathers, 2002). As parent-child contact is seen as a positive factor in enhancing reunification outcomes, where assessed to be feasible contact should take place regularly and in an environment that is conducive to the repair and nurturing of relationships. Parents in this study and previous studies on children entering care (Kapp & Vela, 2004;Spratt & Callan, 2004) experience anger, sadness and loss. The importance of social workers tuning into parents' feelings at an early stage is to be emphasised in laying the groundwork for contact visits and ongoing trustful relationships with parents in the event of reunification or alternative case plans. Social workers play an important part in supporting parents to ensure contact for parents and children is as positive as possible. They also play a significant role in enabling parents and children to express their views and participate in decisions affecting their children.
The relationship between carer and parent was also seen as crucial, and reiterates prior research that sees the establishment of strong carer-family and worker-family relationships as vital (Lewis, 1994;Walton et al., 1993). When the carer and the parent formed a constructive relationship, the carer often modelled effective parenting techniques, which assisted the parents in learning good strategies for responding to their children. The relationship between parents and their caseworker was also seen as important. Caseworkers who met regularly with parents were more likely to gain their trust and were perceived by parents to treat them with greater respect. The findings show how the fostering of positive relationships is the cornerstone of effective reunification practices (Fernandez, 2012). Thus it can be surmised that a relationship premised on trust, mutual respect and negotiated guidance may be vital for enhancing the quality of child welfare decision making (Fernandez, 2007;Ruch, Turney, & Ward, 2010).
Overwhelmingly, both the thematic and statistical data highlight the multiple and intersecting pre-intervention environmental, social and psychological factors that affected families in this study. A majority of families reported that they received social security benefits as their sole source of income, and lived in public housing. From the qualitative data, it is apparent that financial and housing deprivation permeates all aspects of the families' lives, creating numerous interconnected stressors. Substance abuse was prevalent in more than half of families (60.0%), and was linked to issues of child abuse and neglect, domestic violence and mental illness. Nearly a third of families (30.0%) were seen to have parents with serious and untreated mental health issues. Caseworkers highlighted the isolation of birth families. Many parents reported no friends or extended family upon whom they could rely. In addition to striking lack of informal social support, families also lacked connection with any formal support mechanisms until the agency became involved. Caseworkers seldom cited families who entered care simply for one particular difficulty or reason. Rather, multiple factors were often prevalent. This is consistent both with studies in the United States, where multiple and co-occurring problems are often witnessed (Choi & Ryan, 2007;Harwin et al., 2012), as well as in the United Kingdom, where a recent English study of the patterns and outcomes of reunification found multiple adversities experienced by children from the time they first came to be known to children's services up until before entering care (Farmer et al., 2011).
In terms of service interventions caseworkers cited a number of interventions that needed to be implemented in order to improve parenting capacity and child well-being. A majority of parents were referred to multiple services relevant to their particular needs. Continuity in services and the need to encourage some parents to consistently attend such services were seen as vital. Services included parenting education to enhance knowledge about the effects on children of domestic violence, neglect and substance abuse, referral to drug rehabilitation services, financial planning, legal services, counselling, advocacy and housing support. Due to the small size of study sample, service utilisation was investigated using Chi-square tests only. However, the preliminary analyses showed that younger mothers and more educated mothers received and responded to educational services more than their counterparts. It is possible that educational services provided to these mothers contributed to the improvement of NCFAS-R scores and further contributed to the reunification. However, with the limited data, the current study was not able to confirm this. Parenting is one of the many factors that place families at risk for maltreatment. This finding contributes to the expanding body of research on parenting education and its responsiveness to specific parenting groups.
The dearth of services available for birth families to address the problems which eventuate in care placement is commonly cited in research on reunification outcomes. Previous studies highlight the correlation between successful outcomes and service variables, including the availability of concrete services, the creation of positive worker family relationships and the availability of various education and skills training to birth parents (Lewis, 1994;Walton et al., 1993). Other positive correlations include the nature of families' support network, including the quality, reliability and size of this network (Festinger, 1996;Fraser et al., 1996). To reduce the structural risk factors which exacerbate problems for children and their families, it is essential to address the wider social-structural context in which these families live. This entails addressing welfare arrangements, income support, housing, neighbourhood safety, child care and health care. This includes a structured and staged approach to reunification supported with a package of services and comprehensive parenting support. Further, advocacy services that enable parents to access community agencies concerned with housing, income support and health services are crucial (Dakof, Cohen, & Duarte, 2009;Grant et al., 2011). Housing problems are noted to trigger entry to care and delay family reunification. In this context the emphasis on greater collaboration between housing and child welfare systems is to be noted (Farrell, Britner, Guzzardo, & Goodrich, 2010;Harburger & White, 2004). Families with complex needs also require relationshipbased and multipronged approaches to support parents in meeting the challenges of reunification.
The post reunification period is an important period to identify re-emergence of threats/concerns to the child's well-being. Reunification can present a variety of challenges for families particularly when children have been away. Both child and family experience disruptions and discontinuities at entry to care and return. Such discontinuities in family membership can make re-establishment of family relationships a challenge and create additional stress for the family (Rycus & Hughes, 2008). Due to the complexities accompanying reunification families will need intensive supports to sustain child and family well-being both at the time of reunification and for extended periods after reunification. This reinforces importance of careful assessment and post reunification case plans to ensure continuing safety and well-being.

Further research
Research has a critical role to play in more clearly conceptualising and helping delineate reunification as a distinct domain of social work practice, as well as addressing critical gaps in empirical knowledge. There is limited data on post reunification outcomes and this is a fertile area for future research. Most notably, reunification can create new challenges and stressors for families when children return need to be researched. This study, for example, did not address post reunification outcomes and the phenomena of re-entry to care following reunification. Whilst valuable information on post-reunification outcomes (including recurrence of abuse) and re-entry to care are increasingly documented in an emerging body of research (Berrick, 2009;Farmer et al., 2011;Fuller, 2005), this can usefully be expanded. The evaluation of practice approaches and assessment methods that facilitate effective reunification practice can be explored.
Examining the relationship between barriers to reunification and types of services offered to families would assist with planning effective interventions and services. There is also a need for research studies that tap the perspectives of fathers/father figures, which has been relatively neglected, partly due to the difficulties of accessing and engaging suitable samples. Research could be undertaken on the role that carers and caseworkers assume in formal decision making process in reunification matters.