Child and Parent Outcomes in the London Family Drug and Alcohol Court Five Years On: Building on International Evidence

Parental substance misuse is a major social problem of international concern and frequent cause of referral to child protection agencies and court proceedings due to the significant child harm it can cause. Family drug treatment courts have emerged as one of the most promising interventions in recent years to enhance prospects for reunification, substance misuse cessation, and cost savings on out of home care. Despite the better results of Family Drug Treatment Courts at the end of the court case, there has been little investigation of their longer-term outcomes. This article first reviews the international evidence on longer-term outcomes, before presenting new evidence on outcomes of the London Family Drug and Alcohol Court (FDAC), up to five years after the court case ended. The discussion considers the findings of the London FDAC, the first such court in England, in light of the international evidence and makes further research and policy recommendations. The article concludes that the scant international evidence base does not permit an authoritative answer on the contribution of family drug courts to the durability of family reunification and substance misuse cessation. It does however show the need for more family support, particularly in the first two years after court proceedings end. Despite the many challenges, largescale post-intervention evaluations of family drug courts are urgently needed to inform public policy and practice. I . I N T R O D U C T I O N Parental substance misuse is a major and long-standing social problem of international concern. Frequently cited in referrals that bring parents to the attention of child protection services, substance misuse can undermine parenting and result in serious harms to children (Chaffin et al, 1996; Wolock and Magura, 1996; Wekerle et al, 2007; Dawe et al, 2008; Cleaver et al, 2010; Farmer et al, 2011; Forrester and Harwin, 2011; Guy et al, 2012; Park and Schepp, 2015; Velleman and Templeton, VC The Author(s) 2018. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 140 International Journal of Law, Policy and The Family, 2018, 32, 140–169 doi: 10.1093/lawfam/eby006 Advance Access Publication Date: 5 April 2018 Article D ow naded rom http/academ ic.p.com law am /article//140/4962132 by gest on 16 Sptem er 2021 2016). As a leading cause of child neglect internationally, parental substance misuse is substantially over-represented in care proceedings in England and it plays a prominent role in recurrent proceedings (Broadhurst et al, 2017). Children affected by parental substance misuse are also at greater risk of experiencing substance misuse, depression, and a range of health problems as adults (Dube et al, 2003; Anda et al, 2006; Strine et al, 2012). Although parental substance misuse is certainly amenable to treatment, a critical issue for services is that recovery is not straightforward despite the best efforts of parents and professionals. Multiple international studies confirm the difficulties of sustaining positive change for this group of parents, which places their children at heightened risk of reunification breakdown, where they are returned home from public care (Harwin and Ryan, 2008; Farmer et al, 2011; Harwin et al, 2013). For these reasons, there is a strong international imperative to find effective treatments that can produce durable and safe child reunification for families whose prospects for remaining intact are otherwise bleak. Family Drug Treatment Courts (FDTCs) are one of the most promising interventions to have emerged in recent years in efforts to enhance child reunification by tackling parental substance misuse. These courts aim to treat as well as to adjudicate, in the belief that without attention to the problems that underlie parental substance misuse, lasting change cannot be achieved (Winick, 2002; Winick and Wexler, 2015). FDTCs originated in the USA and their development was catalysed by the Adoption and Safe Families Act 1997 (Cooper, 2017) which aimed to promote timely permanency within 12 months of a child entering temporary custody and encouraged concurrent planning of alternative placement options. The primary goal of the family drug court is to find a permanent home for the child whether through reunification, placement with relatives or adoption. These courts have now spread to Australia, England and, most recently, in Northern Ireland. They are only found in adversarial legal systems and the largest number outside the USA is found in England where they are called Family Drug and Alcohol Courts (FDACs). In Australia only one such court has been set up as a 3-year pilot in the Children’s Court of Victoria and the court in Northern Ireland is still at the planning stage. A body of international research offers largely consistent evidence that FDTCs deliver better short term child welfare outcomes for children and families than conventional family courts—assessed at the end of the court process and treatment (Worcel et al, 2007; Oliveros and Kaufman, 2011; Gifford et al, 2014; Harwin et al, 2014; Lloyd, 2015). In the main, evidence derives from the US, where the number of FDTCs totals over 300 (Children and Families Futures, 2013 rev 2015). Positive outcomes are that parents are more likely to enter treatment for drug and alcohol substance misuse at a timely point, to complete treatment and that more children are reunified with parents (Young et al, 2003; Dakof et al, 2009). Children also spend less time in foster care (Marlowe and Carey, 2012; Children and Families Futures, 2013 rev 2015), which delivers cost savings for child welfare services (Marlowe and Carey, 2012; Harwin et al, 2014; Reeder and Whitehead, 2016). A few studies suggest that FDTCs also benefit the psychosocial functioning of the child, adult, and family (Spataro, 2011; Brook et al, 2015; Cosden and Koch, 2015) and their health and reduce new arrests (Carey et al, 2010a,b). The London FDAC 5 years On 141 D ow naded rom http/academ ic.p.com law am /article//140/4962132 by gest on 16 Sptem er 2021 However, where evidence is particularly needed is in the analysis of the longerterm outcomes of the FDTCs (York et al, 2012; Eldred and Gifford, 2016). What happens to parents and children in the years following parents’ graduation from these specialist courts? For example, do parents relapse once the intensive oversight of the treatment phase concludes and the court process ends, or is there a lasting FDTC effect? Such questions are critical to the further development of the FDTCs in terms of practice and policy but in the absence of robust follow-up studies, it is impossible to offer any answers. Taking up the challenge of responding to this gap in our knowledge we bring together the American evidence and findings on the longer term outcomes of the first English FDAC in care proceedings. While the London FDAC has been described in our earlier work (Harwin et al, 2014), the present study provides a unique longitudinal perspective on the London FDAC and is distinctive in its exploration of investigating the sustainability of permanency arrangements for all categories of children whether reunited or not. Although our measures have been adapted for the English context, this work adds to international literature because it highlights the importance of follow-up for families receiving specialist court intervention. At the same time, the study was not without its difficulties in terms of the constraints of small sample sizes, as has been discussed in the broader international FDTC literature (Bruns et al, 2012; Chuang et al, 2012). The article is divided into four further sections. In Section II we provide a comprehensive review of the studies reporting a longer view of the outcomes of the FDTCs, and comment on their methodological strengths and limitations. In Section III we briefly describe FDAC London while in Section IV we report the aims, study design, methodology, and findings of our evaluation. In the final section we return to our central question and consider how new evidence from the London FDAC adds to the international evidence on the longer-term effectiveness of family drug courts, the challenges in conducting follow up studies and consider the implications for practice, policy, and research. I I . S T U D I E S P R O V I D I N G A N E V A L U A T I O N O F T H E L O N G E R T E R M I M P A C T O F T H E F D T C S A comprehensive search for published studies found only a very small number of US studies, reporting analyses of longer-term outcomes, up to four years beyond the treatment phase (see Table 1). As can be seen, it is clear that in contrast to the results at the end of the treatment phase and court process, the evidence on the longer-term outcomes of FDTCs is more equivocal. This finding may in part reflect problems of small sample size, short length of follow-up and control group comparators (Table 1) which make it difficult to draw robust conclusions from the studies. Nevertheless, this small body of work warrants detailed consideration because it shows the difficulties in evaluating a complex intervention such as the FDTC. The studies provide a strong warrant for further studies that examine parent and child welfare outcomes beyond FDTC treatment over the longer term. The first study by Ashford (2004) compared outcomes at 24 months for a small sample of 33 FDTC parents and their 46 children with a group of 42 parents 142 Judith Harwin et al. D ow naded rom http/academ ic.p.com law am /article//140/4962132 by gest on 16 Sptem er 2021


I. INTRODUCTION
Parental substance misuse is a major and long-standing social problem of international concern. Frequently cited in referrals that bring parents to the attention of child protection services, substance misuse can undermine parenting and result in serious harms to children (Chaffin et al, 1996;Wolock and Magura, 1996;Wekerle et al, 2007;Dawe et al, 2008;Cleaver et al, 2010;Farmer et al, 2011;Forrester and Harwin, 2011;Guy et al, 2012;Park and Schepp, 2015;Velleman and Templeton, V C The Author(s) 2018. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 2016). As a leading cause of child neglect internationally, parental substance misuse is substantially over-represented in care proceedings in England and it plays a prominent role in recurrent proceedings . Children affected by parental substance misuse are also at greater risk of experiencing substance misuse, depression, and a range of health problems as adults (Dube et al, 2003;Anda et al, 2006;Strine et al, 2012). Although parental substance misuse is certainly amenable to treatment, a critical issue for services is that recovery is not straightforward despite the best efforts of parents and professionals. Multiple international studies confirm the difficulties of sustaining positive change for this group of parents, which places their children at heightened risk of reunification breakdown, where they are returned home from public care (Harwin and Ryan, 2008;Farmer et al, 2011;Harwin et al, 2013). For these reasons, there is a strong international imperative to find effective treatments that can produce durable and safe child reunification for families whose prospects for remaining intact are otherwise bleak.
Family Drug Treatment Courts (FDTCs) 1 are one of the most promising interventions to have emerged in recent years in efforts to enhance child reunification by tackling parental substance misuse. These courts aim to treat as well as to adjudicate, in the belief that without attention to the problems that underlie parental substance misuse, lasting change cannot be achieved (Winick, 2002;Winick and Wexler, 2015). FDTCs originated in the USA and their development was catalysed by the Adoption and Safe Families Act 1997 (Cooper, 2017) which aimed to promote timely permanency within 12 months of a child entering temporary custody 2 and encouraged concurrent planning of alternative placement options. The primary goal of the family drug court is to find a permanent home for the child whether through reunification, placement with relatives or adoption. These courts have now spread to Australia, England and, most recently, in Northern Ireland. They are only found in adversarial legal systems and the largest number outside the USA is found in England where they are called Family Drug and Alcohol Courts (FDACs). In Australia only one such court has been set up as a 3-year pilot in the Children's Court of Victoria and the court in Northern Ireland is still at the planning stage.
A body of international research offers largely consistent evidence that FDTCs deliver better short term child welfare outcomes for children and families than conventional family courts-assessed at the end of the court process and treatment Oliveros and Kaufman, 2011;Gifford et al, 2014;Harwin et al, 2014;Lloyd, 2015). In the main, evidence derives from the US, where the number of FDTCs totals over 300 (Children andFamilies Futures, 2013 rev 2015).
Positive outcomes are that parents are more likely to enter treatment for drug and alcohol substance misuse at a timely point, to complete treatment and that more children are reunified with parents (Young et al, 2003;Dakof et al, 2009). Children also spend less time in foster care (Marlowe and Carey, 2012;Children andFamilies Futures, 2013 rev 2015), which delivers cost savings for child welfare services (Marlowe and Carey, 2012;Harwin et al, 2014;Reeder and Whitehead, 2016). A few studies suggest that FDTCs also benefit the psychosocial functioning of the child, adult, and family (Spataro, 2011;Brook et al, 2015;Cosden and Koch, 2015) and their health and reduce new arrests (Carey et al, 2010a,b). However, where evidence is particularly needed is in the analysis of the longerterm outcomes of the FDTCs (York et al, 2012;Eldred and Gifford, 2016). What happens to parents and children in the years following parents' graduation from these specialist courts? For example, do parents relapse once the intensive oversight of the treatment phase concludes and the court process ends, or is there a lasting FDTC effect? Such questions are critical to the further development of the FDTCs in terms of practice and policy but in the absence of robust follow-up studies, it is impossible to offer any answers.
Taking up the challenge of responding to this gap in our knowledge we bring together the American evidence and findings on the longer term outcomes of the first English FDAC in care proceedings. While the London FDAC has been described in our earlier work (Harwin et al, 2014), the present study provides a unique longitudinal perspective on the London FDAC and is distinctive in its exploration of investigating the sustainability of permanency arrangements for all categories of children whether reunited or not. Although our measures have been adapted for the English context, this work adds to international literature because it highlights the importance of follow-up for families receiving specialist court intervention. At the same time, the study was not without its difficulties in terms of the constraints of small sample sizes, as has been discussed in the broader international FDTC literature (Bruns et al, 2012;Chuang et al, 2012).
The article is divided into four further sections. In Section II we provide a comprehensive review of the studies reporting a longer view of the outcomes of the FDTCs, and comment on their methodological strengths and limitations. In Section III we briefly describe FDAC London while in Section IV we report the aims, study design, methodology, and findings of our evaluation. In the final section we return to our central question and consider how new evidence from the London FDAC adds to the international evidence on the longer-term effectiveness of family drug courts, the challenges in conducting follow up studies and consider the implications for practice, policy, and research.

II. STUDIES PROVIDING AN EVALUATION OF THE LONGER-TERM
IMPACT OF THE FDTCS A comprehensive search for published studies found only a very small number of US studies, reporting analyses of longer-term outcomes, up to four years beyond the treatment phase (see Table 1). As can be seen, it is clear that in contrast to the results at the end of the treatment phase and court process, the evidence on the longer-term outcomes of FDTCs is more equivocal. This finding may in part reflect problems of small sample size, short length of follow-up and control group comparators (Table 1) which make it difficult to draw robust conclusions from the studies. Nevertheless, this small body of work warrants detailed consideration because it shows the difficulties in evaluating a complex intervention such as the FDTC. The studies provide a strong warrant for further studies that examine parent and child welfare outcomes beyond FDTC treatment over the longer term.
The first study by Ashford (2004)  The programme is for postadjudication/post disposition parents with children aged up to 9 years at entry. Duration of programme (12-18 months a child adopted (5% v 36%). The rate of re-arrests was significantly lower for FATC participants up to 2.5 years only.
No statistically significant differences were found between FATC participants and comparison sample for time spent in foster care, reunification or placement changes.
(51 children) who refused the FDTC, and 45 parents (72 children) who received treatment as usual. The study took place in Pima County, Arizona. The FDTC here is a parallel court where two different judges have jurisdiction of the 'dependency' petition 3 and of the substance misuse treatment. As well as reporting on outcomes at the end of the proceedings, the study included follow-up of re-entry to care within 24 months of the petition. It found no statistically significant differences between FDTC and the 2 comparison groups. These results do not suggest that Pima County Family Drug Court was more successful in preventing re-entry into the care system after treatment completion. Three studies are reported in 2007. Boles et al (2007) adopt the language of recidivism to denote the potential for re-entry to out of home care, for children previously returned home following FDTC intervention. The study compared the outcomes for the FDTC participants with the outcomes for a comparison group whose members did not participate in the FDTC in the county of Sacramento. To minimize demand on county staff, the research team utilized administrative data, linking existing multiple Child Protective Services, Alcohol and Drug Services, and court data systems. Data was provided for both a comparison (111 parents and 173 children) and treatment group (573 parents and 861 children), for whom 24-month outcome data was available. The study controlled for any differences in caregiver baseline characteristics in all analyses of treatment and outcomes.
For all cases, the follow-up period was 24 months from the start of treatment. Regarding follow-up, the researchers distinguished between 'recidivism' and 'reentry'. Recidivism referred to the percentage of children 'who came back into out-ofhome care following a new allegation after their prior case had been closed and where dependency had been terminated' (Boles et al, 2007, p. 169). Re-entry referred to the percentage of children who were reunified with their families during the 24 months following the project start date, but came back into out of home care, before their case was closed. The authors reported that they found no statistical differences in recidivism rates between the comparison and treatment groups. The overall rate of recidivism for both groups was extremely low, with less than 1.5% of children experiencing recidivism during the study period (1.7% in the FDTC group and 0% in the comparison group). The authors also reported no statistical differences in the rates of re-entry for children to out of home care, with 21.7% of the children experiencing re-entry following reunification. In the overwhelming majority of cases the reason was parental relapse. As with Ashford, from these findings, it does not appear that the FDTC delivered better outcomes for children in the longer-term, than 'business as usual'.
Another study in 2007 evaluating Maine's state wide FDTC programme (Zeller et al, 2007) followed Boles et al (2007) in its distinction between investigating reentry prior to case closure and child welfare recidivism, defined as a new Department of Health and Human Services 4 case opening within one year of the date of final case closure. FDTC cases (n ¼ 49) were compared with cases heard from 3 comparison jurisdictions (n ¼ 38) and with a sample of substance misusing parents known to child welfare services prior to the creation of the FDTC (n ¼ 55). The researchers reported few differences between the family drug court and the two comparison groups in child welfare recidivism outcomes. Because the recidivism rate was so low in all three groups, it was not possible to test for statistical significance. However, child welfare recidivism among family drug court participants in this study (6.8%) was lower than for family drug court programmes nationally (14-23%). Commenting on possible reasons for these variations, the authors point out that the length of follow-up was likely to affect rates of child maltreatment recurrence as would the definition of maltreatment used in the different studies. This was an important observation but for the purposes of this article, the most important finding here is that FDTC did not demonstrate better outcomes than the comparison groups.
The next study in 2007 by Worcel et al (2007) is particularly interesting because it is a national largescale evaluation using a quasi-experimental design and includes a wide range of outcome variables to measure recidivism. The sample comprised 739 FTDC mothers and 1120 comparison mothers with cases drawn from 4 main study sites. Differences in the particular FDTC models operating in the different sites were noted and accounted for in the study design. To measure recidivism the study examined six main child and parent measures. They comprised substantiated referral to Child Protective Services (CPS), new court proceedings after the original case had concluded, termination of parental rights, subsequent out-of-home placements, and the birth of subsequent babies born affected by drugs. None of these measures generated statistically significant differences between the two groups of mothers across the four sites, but in two sites FDTC mothers had fewer babies born withdrawing from drugs and in one site more FTDC mothers had a subsequent referral to CPS. As with the study by Boles et al, the inability of this study to detect recidivism is considered to be linked to the very short follow up window, only 2 years from the start of the petition, and with 61% of cases still in the FTDC programme at the end point. While recommending the need for much longer follow up on all these measures for larger numbers, the authors note that recidivism was more frequent when the original case involved termination of parental rights and children were not reunified.
A study by Twomey et al (2010) carried out in Rhode Island is noteworthy because of its focus on perinatal substance misusers, its use of a wide range of internationally recognized standardized measures to report on maternal and infant developmental outcomes, and its findings. It did not however have a comparison group. 52 mothers and their 54 infants were tracked until the children were aged 30 months to report on maternal and infant developmental outcomes and placement permanency. Infant assessments were carried out at 6 monthly interviews from 12 to 30 months with a developmental assessment at 30 months including an attachment interview if the child was still living with its mother. The FDTC provided data on whether the mother had graduated, was still in the court process, or it had terminated. The study found evidence of deterioration between follow-up at 12 and 24 months in most maternal outcome measures. These included substance misuse, mental health symptomatology, inappropriate or negative parenting attitudes, and parenting stress. All of these results were statistically significant. 5 As regards child development, some new concerns in respect of infant development were found by 30 months: deterioration in pervasive developmental delay was statistically significant. Since over 75% of biological mothers and children lived together at each time point the results are particularly noteworthy.
The authors point out that their study had two main limitations. Apart from the small sample, the Rhode Island FDTC also includes an enhanced support programme for the mothers, which could affect the lessons that can be drawn for other FDTCs, and as noted, there was no comparison group. The findings however merit attention because they are based on robust measures which point to the importance of addressing the long-term needs of these families 'even when promising interventions like FDTC are used' (Twomey et al, 2010, p. 23).
Turning to two of the most recent studies identified in our review, Bruns et al (2012) focused on fresh allegations of child abuse for families undertaking treatment in an integrated FDTC (in which the same judge presided over the dependency hearing and the substance misuse treatment). The study relied on administrative records from 4 different sources which were linked by using 'indirect identifiers'. Regarding follow-up, which was measured in terms of fresh allegations and allegations that were founded within 24 months of the start of the court proceedings, no statistical difference was found between treatment and control groups. The authors suggest that parents in the FDTC were more likely to have children returned to them, hence their likelihood of reinvestigation was higher. To address this, the researchers re-analyzed the data by selecting only those parents who had at least one of their children returned to them. Based on re-analysis of the data in this way, the authors found that parents in the FDTCs were about half as likely as comparison parents to have a subsequent investigation (33% versus 64%) but these results were not statistically significant. 6 The authors make the point that the re-analysis resulted in further reduction in sample size which can make it difficult to detect anything other than large treatment effects. This is an important point which underscores the importance of replicating the Bruns et al (2012) study on the basis of larger sample sizes. The work of Bruns et al (2012) clearly illustrates the methodological challenges in evaluating complex interventions such as the FDTC, which often yield only small treatment populations for analysis.
Turning to our second study in 2012 by Chuang et al (2012) the authors reported the findings from a study of an integrated FDTC operating in a single US county (Hillsborough County, FL). The researchers examined the impact of the FDTC on: reunification, time to permanency and re-entry to care. Reporting only on the latter child welfare outcome given our focus on longer-term consequences for children, the authors found that re-entry rates were very low for children 12 months after reunification with their families (2% of the 53% of children who were reunified with parents as an outcome of FDTC). These re-entry rates were lower than for the comparison group, where 12% of children re-entered care within 12 months of reunification, and the difference was statistically significant. 7 The authors also quote findings from the multi-state foster care data archive produced by Wulczyn et al (2000), which was based on a far larger population of families across multiple jurisdictions and found re-entry rates much higher in the US at 21-28%. In accounting for the discrepancy in findings, Chuang et al (2012) attribute these to differences in models of FDTCs. Boles et al (2007) examined a 'parallel' model, whereas Chuang et al (2012) examined the 'integrated' FDTC. The particular nature of FDTCs must be factored into any analysis.
The final study we consider in this review (Mackin et al, 2013) explores the issue of re-arrests as well as measures considered by other studies on placement outcomes and termination of parental rights. Reviewing the data from 2 family drug court sites in Marion and Jackson County, the study followed up a sample of post-adjudication/ post-disposition participants with children under the age of 9 who entered the programme between January 2006 and 2008 and followed them up in 2013 with a matched comparison sample who with, the exception of race were not different from the family drug court participants-a factor that was controlled for in all the outcome analyses. The longer follow up (between 2 and 4 years following the original petition date or entry date to the programme) was particularly useful in demonstrating longer term outcomes. Four years after programme entry, participants in the family drug court were significantly less likely to experience termination of parental rights and their children were significantly less likely to be adopted. Criminal justice outcomes were also superior. Programme participants were significantly less likely to be rearrested for drug related offending at one year following programme entry and to be rearrested for any type of offence in the first 2.5 years. 8 Between group differences in re-offending were not significant after that point. However, a limitation of the study was its small sample size (39 family drug court participants and 49 comparison cases).
To summarize, on the basis of published evidence to-date (8 studies of varying size and quality), it is difficult to conclude that the FDTC has a greater protective impact in terms of insulating children from re-entry to out of home care following reunification. However, recidivism rates are low for both treatment and control groups, suggesting that FDTCs do not have an adverse effect on child welfare outcomes in the longer-term. As Green et al (2007) and Bruns et al (2012) have also described, it may be that the greater surveillance of FDTC families beyond treatment may predispose children in these families to higher rates of CPS allegations. But the low rates of recidivism may also reflect the short follow up period after the court supervision ended. With regard to scope, the prime concern has been about the durability of reunification. Long-term follow up of the impact of FTDCs on all entrants to include those who were not reunified is not explored. In short, it will be important to carry out further largescale evaluations of FDTCs to establish their long-term impacts.

III. THE LONDON FDAC
The London FDAC was set up in 2008 with part funding from central government to help achieve its first ever family drug policy strategy (HM Government, 2008) and to improve child welfare outcomes for children taken into public care (Department for Education andSkills, 2006 and. With research showing in one London survey that cases involving parental substance misuse accounted for 62% of care proceedings (Forrester and Harwin, 2006), government supported the view that it was time for a new approach. Spearheaded by Judge Nick Crichton, and drawing on the US evidence of the short term success of American FDTCs at the end of the court case and concerns over poor communication between substance misuse services and children's services in England, the London FDAC was launched. It was piloted between January 2008 and April 2012 in the Central London Family Proceedings Court, following adaptation to fit within the English and Welsh legal and welfare systems (Harwin and Ryan, 2008). Three pilot authorities match-funded the government's financial support and an independent evaluation was introduced from the outset. The Tavistock Portman National Health Trust was responsible for the governance of FDAC.
FDAC provides parents with an opportunity to demonstrate capacity to change by providing intensive supervision of families and tighter co-ordination of service inputs, coupled with continuity of judicial authority. Local authorities refer cases into FDAC when they issue care proceedings under section 31 of the Children Act 1989, where the main concern of the local authority is that the child is suffering actual or likely significant harm as a result of parental substance misuse. It is important to emphasize that participation in FDAC is voluntary in the belief that this choice will help increase parental engagement. If parents decide they do not wish to have their case heard in FDAC, it is heard in ordinary care proceedings.
The London FDAC follows the model of an American 'integrated FDTC' whereby the same judge has jurisdiction for both the care proceedings and FDAC treatment intervention. The main features of FDAC are judicial continuity, fortnightly judge-led review hearings without lawyers present, and a specialist multidisciplinary team (MDT). Independent of the local authority, the MDT advises the court and provides intensive treatment and support to parents as well as close monitoring of their progress. The specialist team also links families with local community substance misuse and family support services. The non-lawyer review hearings are the court-based therapeutic forum for the problem-solving component of FDAC. Also central to FDAC is the role of the specialist team in problem solving with the Judge, and with families and practitioners outside of the court room.
All these features make FDAC radically different from ordinary care proceedings. In ordinary care proceedings in England there is no independent multidisciplinary team or judge-led review hearings where the judge plays a problem-solving role and actively seeks to motivate parents to change. Nor do parents engage in conversation with the judge. Instead their views are presented by their legal representative to the judge. Unlike some American FDTCs, the main sanction is that if parents do not comply with their FDAC plan, the case reverts to mainstream public law care proceedings where the local authority still has to prove to the court that the child is at unacceptable risk. At this point FDAC ceases to have any further involvement. All planning for alternative care becomes the responsibility of the local authority. ). At a time of austerity this was an ambitious plan. However a new opportunity arose in 2014/15 for local authorities to develop FDACs in partnership with the courts. The government launched the Children's Social Care Innovation Fund, a major competitive tender to encourage the testing of promising initiatives to provide evidence on their potential for scale-up, replication, and durability. FDAC was successful, and with this funding, 13 new FDACs teams were established in 16 courts, serving 21 local authorities in 2015-2016 to be supported by the creation of an entirely new body, the FDAC National Unit. All the new FDACs were expected to follow the approach of the London FDAC to ensure that their outcomes were not affected by a different delivery model. More recently, FDAC has received £6.2 million over 7 years from the government Life Chances Fund, with the aim of benefiting 2,400 individuals. The scheme is expected to attract support from some local authorities which have not previously offered FDAC as well as those which currently provide the intervention. The scheme operates on a payment by results model.

The wider roll-out of FDAC
The roll-out of recent FDACs takes place in a new legal framework following the passage of the 2014 Children and Families Act. Amongst its many measures, of particular relevance to FDAC is the introduction of a statutory 26 weeks for normal completion of the proceedings, compared to an average length of 12 months in 2008. 9 In exceptional circumstances applications can be made to extend the proceedings and because it can be so difficult to achieve abstinence in 26 weeks FDAC is recognized as an exemption to the rule provided there is evidence to show motivation to change, ability to maintain it, and that change takes place within the child's timescales. 10

IV. LONGER-TERM EVALUATION OF FDAC
The decision to examine the longer-term outcomes of the London FDAC was prompted by the roll out of FDAC to other regions of England. Moreover, and of particular relevance to this article, the incidence of neglect in reunified FDAC families in the 2014 evaluation was significantly lower 11 one year post-treatment compared to families who went through ordinary court and services, albeit based on analysis of a very small number of cases (Harwin et al, 2014). For all these reasons, there was considerable national interest and impetus for further evaluation of the longer-term outcomes of FDACs. The evaluation study, After FDAC: outcomes 5 years later (Harwin et al, 2016), was funded for one year by the Department of Education's Children's Social Care Innovation Programme.

Research objectives and questions
The over-arching objective of the study was to produce evidence about the longerterm impact of FDAC in order to inform the wider roll-out in England and decisions about future investment. Two main questions framed the evaluation study. The first was to compare the durability of safe family reunification and maternal substance misuse cessation in FDAC and comparison cases. The second was to establish if there were any differences between non-reunified FDAC and comparison children in the durability of their out of home placement outcomes. The rationale for this The London FDAC 5 years On 151 second line of inquiry was to help establish if out of home care, despite its initial higher costs than reunification, in the longer term, gave children a better chance of putting down new roots and had cost benefits.

A. Methodology
The study used a quasi-experimental non-randomized design and the cohort was tracked at three time points-at the start and end of the care proceedings and for a maximum of five years after proceedings ended ( Table 2).
The FDAC cohort comprised all cases that entered the London FDAC between January 2008 and August 2012 and they were tracked for up to 5 years after the care proceedings ended and the FDAC intervention stopped. They were compared with cases from three comparison local authorities which did not provide FDAC. These cases were heard in ordinary care proceedings in the same Family Proceedings Court as FDAC. In all cases parental substance misuse was a key factor in the local authority's decision to issue section 31 care proceedings on the grounds that the child was subject to actual or likely significant harm. In line with FDAC policy, cases were excluded from the cohort if the parent was experiencing florid psychosis as informed consent to join the intervention would not be possible. Serious current domestic abuse posing a risk to child safety and a history of severe parental physical or sexual abuse were also grounds for case exclusion. These criteria were also used by the comparison local authorities in selecting their cases for the study.
Profiling was carried out on the cohort at the start of the proceedings to establish whether there were sufficient similarities to merit comparison of end of proceedings and follow-up outcomes. To this end, socio-demographic and psychosocial data about FDAC and comparison mothers and their children and information on the maternal difficulties that had led to the proceedings were compared. A similar exercise was carried out in respect of the children.

B. Data sources and management
Data was extracted from administrative records. They included the child's electronic case file held by local authority children's services and the Cafcass national electronic One FDAC mother and two comparison mothers died during the course of the proceedings b Some mothers, 6 FDAC and 1 comparison, are in both the reunification and the non-reunification groups because they were reunited with some children and other children were placed away from them.
case management database which holds information on all public and private law proceedings and legal orders in England in which Cafcass plays a role. The Cafcass electronic database was used to track all cases that returned to court up to July 2014 and to corroborate court related data extracted from the local authority files. The research database used in the 2014 evaluation comprised court file records of FDAC and comparison cases, FDAC files held by the Tavistock Coram NHS trust and local authority children's services files. A Microsoft Access relational research database was constructed for this study and allowed the research team to vary the unit of analysis, according to the research question. It contained quantitative information relating to the FDAC and comparison mothers, the children, their legal outcomes and placement arrangements at the three study time-points. To help shed light on the quantitative follow-up outcome data, qualitative case commentary was constructed by the researchers as they read the local authority children's file and the Cafcass database.

C. Data analysis
To meet the study aims, the relational database was sub-divided into two subsets for the follow-up-a reunification cohort and an out of home cohort. The reunification cohort included all cases in which the mother was reunited with at least one child, to reflect maternal capacity to change, thereby maximizing the size of the 'per mother' reunification sample and reflecting the fact that children from the same family might have different placement needs. The quantitative analysis of maternal substance misuse cessation, family reunification and out of home placement permanency at the end of proceedings in the FDAC and comparison cases is based on cross-tabulated frequencies/percentages. Statistical significance was tested using the Chi-Square test based on calculating the probability of error at the minimum accepted level (p < 0.05). The analyses were carried out using Statistical Package for the Social Sciences.
Results regarding the durability and safety of family reunification, maternal substance misuse cessation and placement sustainability in FDAC, and comparison cases at 5 years after the end of proceedings are based on survival analysis (Clark et al, 2003;Collett, 2015). This statistical approach which calculates the probability of an event such as substance misuse occurring and its timing, was considered to be particularly useful for the study. Its main advantage is that it takes into account and can adjust for varying lengths of follow-up (Lovric, 2011) through estimations. This was a particularly important consideration because follow-up data depended on when the care proceedings had ended and cases that concluded more recently had a shorter follow-up period. Cox regression was used to calculate the proportional hazard of an event occurring and had the advantage of enabling us to include censored data whereby the event of interest has already occurred (eg the subject has withdrawn from the study) or the event (eg substance misuse recurrence) did not take place within the follow up timeframes.
The events we tracked in the follow-up regarding the mothers were cessation of misuse and retaining care of at least one child. However, because of our earlier work and the international literature positing a close relationship between substance misuse and other psychosocial difficulties (Cleaver et al, 2010), we also tracked domestic The London FDAC 5 years On 153 violence and mental health problems, the birth of a subsequent baby and maternal return to court in the follow-up. In all cases we were interested in when the event first occurred in the follow-up as this information also has practice and policy implications. For the comparison of the children, the key events of interest were return to court, experience of further neglect or abuse and whether there had been a change of placement after they had reached their permanent placement. Brief details of each event were logged with accompanying contextual information wherever available.
All the events outlined above chart problems and were based on a single variable. We were however keen to also capture positive progress and to seek a proxy of a good outcome. A mother's outcome was defined as good if none of a combination of the following three key events occurred during the follow-up period: substance misuse, a permanent placement change for a child or children, or return to court. A child's outcome was defined as good if none of the three events occurred in the follow-up: permanent placement change, subsequent neglect, and return to court for new care proceedings. Each of these analyses was carried out per mother and per child.
All results are therefore based on the timing of the first event and all percentages are cumulative, thus providing estimates over the 5-year period (and, exceptionally, for 3 years instead). Kaplan-Meier analyses were used to generate the survival analysis curves and the distributions of the FDAC and comparison groups were tested using the log-rank test (see Appendix 2 of the 2016 report for further details). Table 3 presents the demographic and psychosocial profiles of the mothers at the start of the proceedings (baseline). The similarities between the two samples were striking. As can be seen, over two-thirds of the mothers were aged 30 or older. The misuse of both illegal drugs and alcohol was the most frequent pattern; maternal alcohol misuse alone was least common. In line with the international literature, maternal substance misuse was rarely the only serious psychosocial difficulty: co-occurring mental health difficulties and the experience of domestic abuse were also common. Reinforcing the evidence of the continuities between childhood and adult adversity in this very vulnerable population, a notable proportion of the mothers had previously had a child removed in care proceedings (Harwin et al, 2014;Broadhurst et al, 2015) and had themselves been in care as children . A particularly concerning finding was the discovery that the largest proportion of the mothers in both samples had been known to children's services for more than 10 years-though not necessarily continuously. It raises questions as to whether more could have been done earlier to arrest the problematic substance misuse and reduce the likelihood of the need for care proceedings. In line with our national analysis of care proceedings  a majority of care proceeding cases in the 2008-2012 period concerned only one child (Masson et al, 2008;Broadhurst et al, 2017). B. A descriptive profile of the children at the start of the proceedings Care proceedings are always only brought for the most vulnerable children and the profiling of these cases demonstrated the many ways in which these children were particularly vulnerable (Table 4). Infants under the age of one comprised the largest group (38.3%), a higher proportion than reported in Masson et al, (2008) (28.5%) dealing with all types of care proceedings rather than those triggered by parental substance misuse. This is consistent with the findings from our national study of care proceedings: between 2008 and 2012 infants under the age of one comprised between 29% and 32%. While the overwhelming majority of FDAC cases were brought because of neglect, in line with other studies of care proceedings, in both samples the harm also included physical and emotional types. A striking finding was the extent of health difficulties. They affected nearly half the sample and all age groups. The health difficulties included issues such as neonatal abstinence syndrome, hearing and sight problems, heart murmurs and asthma. Just under a third were described in Table 3. Profiles of the mothers at the start of the proceedings

A. A descriptive profile of the mothers at the start of the proceedings
the files as having emotional and behavioural problems. Amongst the younger children, the problems included bedwetting, hyperactivity, and withdrawn or attentionseeking behaviour. For older children, the problems included lack of self-confidence, difficult behaviour at home or school, and running away from home or school. The results showed that the similarities outweighed the differences in most respects. However statistically significant differences were found in relation to maternal and child ethnicity where there were higher proportions of White mothers 12 in the FDAC cases (73% versus 59%) and proportionately more White children (52.6% versus 39.9%) than all other categories. Chi squared and log rank tests were carried out which established that ethnicity did not act as a confounder of the results at the end of proceedings or at follow-up. (Full profiling details are provided of the cohort below). It is likely that this statistical difference was an artefact of the ethnicity differences in the profiles of the populations in the FDAC and comparison cases. Statistically significant differences between FDAC and comparison cases were also found in relation to cases involving the likelihood of harm, babies born withdrawing from drugs, and the higher proportion of FDAC local authorities plans for placement with family and friends while comparison authorities had a higher proportion of placements with foster carers. Using the same statistical tests as for ethnicity, it was found these differences between the samples did not act as confounders. However it was not possible to establish if comparison families would have accepted the FDAC intervention.
C. Outcomes at the end of the care proceedings There were two statistically significant results. Firstly, a significantly higher proportion of FDAC mothers had ceased to misuse substances (46% versus 30%). 13 This included mothers who were abstaining from alcohol or illegal drugs and those who were stabilized on an agreed treatment programme (such as a methadone script) and were not taking any non-prescribed or illegal street drugs ( Figure 1). Secondly, a significantly higher proportion of FDAC than comparison families were reunited or continued to live together (37% versus 25%) 14 at the end of the proceedings ( Figure 2).
However, there was no difference between the two samples in the time it took children to be placed permanently or in the overall proportion of children with a permanent placement. Three quarters of all the children were living in permanent placements and this included the children who returned to live with their mothers as well as those who were placed in alternative care. When children were reunified, the majority in both samples were made subject to a supervision order (96% versus 98%). 15 This legal order places a duty upon the local authority to advise, assist, and befriend the child for a specified period up to a maximum of 3 years. Most of the supervision orders were made for one year in both samples.
D. The durability of substance misuse cessation and family reunification at followup: estimations from survival analysis There were two statistically significant results. A significantly higher proportion of FDAC than comparison reunification mothers (58% versus 24%) 16 were estimated The London FDAC 5 years On 157 to sustain cessation over the 5-year follow-up. This analysis was based on the number of mothers who had stopped misusing at the end of the proceedings. As can be seen from Figure 3, the maximum period of risk was in the first 2 years after reunification in both FDAC and comparison cases, but thereafter the gap widened and significantly more comparison mothers were estimated to experience substance misuse difficulties.
With regard to the durability of family reunification, a significantly higher proportion of FDAC than comparison mothers who had been reunited with their children at the end of proceedings were estimated to experience no disruption to family stability at 3-year follow-up (51% versus 22%) 17 ( Figure 4). As explained earlier, this way of analysing family stability examined relapse, placement change and return to court as a single composite measure and served as a proxy of a good outcome. It was Information missing on 6 FDAC and 2 comparison mothers. They were all mothers who were not reunited with their children.  only possible to follow-up mothers on this measure for 3 years because thereafter the information became too sparse to report on. Figure 4 illustrates the widening gap over the 3 years post-proceedings between FDAC and comparison mothers regarding the proportion of mothers who did not experience any of the three events. It also shows that the critical period for maximum risk was the first 2 years. The absence of new substance misuse problems was a major factor in the likelihood of a good outcome on this measure. Two further findings in relation to the durability of reunification showed sizeable percentage differences between FDAC and comparison cases but did not reach statistical significance. A higher proportion of FDAC than comparison reunified children were estimated to experience no disruption in the 3-year period after proceedings ended (57% versus 39%). 18 No disruption was defined as a combination of no permanent placement change, no subsequent neglect, and no return to court for new proceedings. Secondly, in relation to the single variable of return to court, a lower proportion of FDAC than comparison reunified children were estimated to start new proceedings due to actual or likely significant harm in the follow-up period (34% versus 55%). 19 No other statistically significant differences between FDAC and comparison cases were found in relation to family reunification. Over the 5-year period around a quarter of all reunified mothers were estimated to experience domestic violence or mental health issues in the follow up. Around one-fifth of FDAC and over one-third of comparison children were estimated to experience neglect when measured as a single variable. The risk was highest in the first 2 years after proceedings ended and in both samples, it was spread across all age bands but no children aged under one were exposed to neglect. Nor were there any episodes of physical or sexual abuse in the follow-up period. Cases where neglect occurred were very likely to return to court but substance misuse was not involved in all neglect cases. Approximately one third of the children in both samples were estimated to experience permanent placement change.
The brief notes kept to record the context to events showed that there was a clear inter-relationship between child and maternal events and their consequences. Mothers in both samples who experienced three or more events such as relapse, domestic violence, and offending in the follow-up period were very likely to have their case return to court for new proceedings. Children's emotional and behaviour problems were linked in all cases with other events, including neglect, relapse, mental health problems, domestic violence and/or the birth of a baby in the follow-up period. Finally, around a third of all children who were reunited with their mother at the end of the care proceedings were estimated to display emotional and behavioural problems in the follow-up period. 20 The problems included uncontrollable and violent behaviour, offending, anxiety symptoms, bedwetting and self-harm. Some difficulties developed during the follow-up period: others were a continuation of preexisting problems.
E. Out of home children: estimates of stability and disruption during the follow-up As noted in the introduction, an important feature of the evaluation was the investigation of outcomes for children who were placed out of home at the end of the care proceedings. A main finding to emerge was that once proceedings ended, there were no statistical differences in the permanency trajectories as between FDAC and comparison cases. The majority of all children (82% versus 78%) 21 were estimated to remain in their permanent placement throughout the follow-up and to experience no disruption. This was the case for each of the main alternative placement types.
However, around a fifth of all FDAC and comparison children were estimated to experience a placement change after they had reached their permanent placement. The proportion was similar for both samples and children aged between 5 and 17 years were most likely to experience placement change. The qualitative analysis showed that the change was nearly always the result of placement breakdown, often linked to the behaviour of the child. It included children absconding from placements and going missing, being aggressive with other children or to their carer, and refusing to go to school. With older children, there were also examples of self-harm and/or attempted suicide. Behaviours that were linked to placement breakdown amongst the younger children were frequently associated with developmental delay. All children who experienced more than one placement change after reaching permanency had serious emotional and behavioural difficulties or mental health problems. Placement breakdowns occurred with and without return to court. Ten per cent of all the children who were placed out of home had their case returned to the court for fresh care proceedings. They included children placed under a Special Guardianship Order and Residence Orders/Child arrangements order. It was not possible to discern any pattern as to the reasons for a return to court.
F. Comparing patterns of placement stability across all placement types There were statistically significant differences in the estimated rate of change of placement for the different placement types. Children placed in long-term foster care had the highest risk of placement change (51%), followed by children placed at home (33%). The estimated rates for children placed with relatives and with fathers were 14% and 12%, respectively. The lowest risk was for children placed with adopters (2%). This pattern applied equally to FDAC and comparison cases. There was a strong association between the age of the child and placement change ( Figure 5).

V. DISCUSSION AND CONCLUSIONS
One of the aims of this article was to draw together the international studies to examine how far FDTCs demonstrate lasting impacts after the family drug court intervention and court process ends. A first discovery from this review is that all the literature is American, and in addition comprises just eight studies over a 9-year period, with nothing published since 2013. The reasons for this are likely to be difficulties in recruiting families into the studies, obtaining funding as well as the The London FDAC 5 years On 161 methodological difficulties that we discuss below but may also reflect a wider problem of lack of recent significant evaluations of FDTCs at the end of the original court case (Cooper, 2017). The consequence is that it is not possible to provide an authoritative answer as to the contribution of specialist family drug courts to the longer term durability of safe reunification and substance misuse cessation. The scant evidence contrasts with the richer evidence base available in relation to adult drug courts in the criminal justice sector (Rossman et al, 2011;Bowen and Whitehead, 2015;Jewell et al, 2017). Here medium and largescale follow up studies in the USA and Canada 22 have been able to explore not only whether there are longer term effects and their nature, but to analyse a range of mediating factors. Evidence from the field of criminal justice indicates that despite the challenges, it is possible to undertake longer term evaluation, affording opportunities for theory building. Given this comparison, research on longer term outcomes of family drug courts can only be considered as in its infancy.
Turning to the London FDAC, our evaluation of the longer-term impact is particularly interesting because it found some evidence of a positive longer-term effect of FDAC. This is in keeping with some of the more recent American studies. Rates of estimated substance misuse cessation were significantly higher in the FDAC cohort than in comparison cases and disruption to family reunification was lower. Moreover the follow up period was longer than in the American studies, providing new evidence that positive effects can endure for at least 3 years for reunification and up to 5 years for cessation.
Notwithstanding the methodological differences that make straightforward comparison with the American studies difficult, the English results merit attention because of their policy and practice relevance and cost implications. But they need to be seen as a starting point for further work because it was not possible to explain the mechanisms that might account for the better FDAC outcomes and who was most (and least) likely to benefit from the programme. No predictive factors emerged to indicate which parents might sustain cessation and which families avoided reunification breakdown at follow-up. Effective evaluation and implementation science demand a theory of change and the ability to say who the programme works for and in what circumstances.
Putting together the findings from the London FDAC and the international evidence, a clear message is that the challenges of recovery indicate the need for more family support to enhance the durability of reunification for all children affected by parental substance misuse. Both the FDAC evaluation and the study by Twomey and colleagues on perinatal mothers and their infants provide similar and complementary evidence. They showed that the maximum risk period for recurrence of risky parenting is in the first two years following the end of proceedings for domestic abuse and mental health difficulties as well as substance misuse. Both studies come to similar conclusions that the level of support needed would vary, according to individual needs, but could for many parents be a useful way of preventing problems from accumulating and getting a hold. This offer of support would build on the research evidence that recovery is a fluctuating and lengthy process that requires different types and levels of support, and that children need good support when they return home from care. It is hard to disagree with the conclusion of Twomey and colleagues 'Conceptualizing permanency as an ongoing state rather than an event with a finite end can normalize interventions for families who would benefit from periodic or more intensive attention, make parenting a less isolated undertaking, and promote more favorable child developmental outcomes' (Twomey et al, 2010, p. 37). Nevertheless in England austerity since 2008 has seen significant cutbacks in early help and family support services, exposing more children to risk of substance misuse (National Children's Bureau [NCB], 2017) while demand for services rises (National Audit Office, 2016).
This analysis of the American and English studies has also generated learning for future family drug court longer term evaluation research. The main interest in the American studies has been on long-term outcomes of reunification but the inclusion of outcomes for non-reunified children in the English study raises important issues which could have wider relevance. The fact that the majority of non-reunified children were less likely to experience placement change or return to court than children who were reunited with parents provides further evidence that family reunification is a more risky option than out of home care. It thereby raises the difficult question of whether family drug courts only postpone decisions about best ways to achieve the long term interests of children. While this debate goes far wider than evaluation, it is nevertheless evaluation studies that can help provide the long-term evidence.
Finally, we are able to confirm common themes regarding the challenges of undertaking evaluation research. The lack of longitudinal follow-up studies reflects a broader problem in child welfare intervention research, rather than a problem specific to FDTC studies (McGhee et al, 2015). Nevertheless these problems present major challenges to FDTC evaluations. The mechanisms that underlie positive or negative outcomes are not easily disentangled, nor can confounders be readily identified or ruled out. Bruns et al (2012) make the important point that 'it is not easy to reliably measure child health and mental health and parent-child attachment over time' for example. An enduring issue is that data based on small sample sizes is typically underpowered and evaluators risk type II errors (failure to reject the null hypothesis) because only large treatment effects can be detected. The reliance on retrospective administrative data for follow-up studies is a further limitation, but the cost of collecting new primary data can be beyond budgets. Thus, on the basis of this international evidence, it is clear that greater investment is needed in the evaluation of longer-term child welfare outcomes of the FDTCs, taking into account differences in treatment models and learning from work that has already been published. Some of the methodological challenges may be difficult to overcome, but the pursuit of an accumulated body of knowledge on the longer-term outcomes of family drug courts is absolutely critical to inform policy and practice.
As Wexler reminds us (Wexler, 2014 and, shifts in the legal framework can create either positive or more hostile environments for problem solving courts such as family drug courts. The 1997 Adoption and Safe Families Act assisted their introduction in the USA with its emphasis on timely permanency planning and prevention of drift and delay. A very similar context also created fertile soil for the introduction of FDAC in England. Yet by 2014, a new legal and policy landscape emerged in England, with greater emphasis on pre-court work, tighter timescales for case completion and monitoring of judicial performance on case completion. While this is a more challenging environment for FDAC, it has nevertheless expanded since the Children and Families Act 2014 and remains part of government's latest drug strategy . But the shifts in law reinforce the critical importance of long-term robust largescale evaluation studies into family drug courts which remain outside the mainstream in both the USA and UK.